5th meeting, 33rd Session of the Committee of Experts on International Cooperation in Tax Matters

If everybody can take their seats, that would be good. Good morning, everyone, and welcome back to this fifth meeting in the 33rd session. I hope everybody had a good night. I'm sorry I wasn't able to join you earlier, but I'm very happy to be here. And Matthew has kept us on schedule. I was able to follow a little bit online. And at one point, I was a little bit worried that we were lagging behind, but we are now somewhat back to schedule. I would like to thank the Swiss government for hosting the reception yesterday. I'm sure it was very good, as per usual. Thank you very much. So we'll proceed this morning. And we will start with a discussion of item 3A, issues related to the United Nations Model Tax Convention. And the basis for the discussion is the Co-Coordinator's Report CRP 12. Now the Committee will receive an update on the work streams and you will be invited to provide comments and guidance on the Subcommittee's initial draft of the consequential changes to the commentary for the inclusion of the subject to tax rule as it's set out in the annex to CRP 12. So I will now invite co-coordinators Vanjeri Kiaria and Antoine Guichet to present the co-coordinators report. And we will also have a presentation from one of the subcommittee participants and committee members on the SDTR paper as contained in the annex, Mr. Eymen Odeh. So please go ahead.

Thank you so much, Trude, and welcome to the meeting. Good morning, everyone. Today we'll start us off with discussions on the subcommittee to update the UN model. Our agenda for today's meeting, if we can have the slides up. So we'll start us off with a recap on the subcommittee's mandate, as all the other subcommittees have been doing, which I've been finding very helpful. And then we'll recap the work streams that are handling all the work under the UN Model Subcommittee, as well as the timelines within which the work will be done. And then we'll have a presentation from our lead drafter in work stream one, which will be a committee member, Eamon, who will present a paper on subsequential changes to the STTR. And then we'll look at the questions that we are putting to the committee and to the floor for discussion. Next slide, please. So a recap of the mandate of the UN Model Subcommittee. We are supposed to provide drafting for the next UN model update, focusing again on issues of the greatest relevance to developing countries, as all the other subcommittees have been providing. What we are also required to do is to identify any linkages between the various subcommittees. and to liaise with the other relevant subcommittees as appropriate to ensure that there is going to be coherence across the committee's outputs. We have been doing this specifically with the transfer pricing and the DGE subcommittees, and we discussed those papers yesterday. That is collaboration that we have already been having and will continue to have with other relevant subcommittees. Our first deadline is for April 2027, which will be the 34th session. During that session, we are supposed to recommend whether a 2029 model update should be produced. We also have another timeline of October 2028, when we are required to make update-to-update recommendations, no later than the 37th session. That is the session within which we should be able to provide the update recommendations. During the previous session, we presented our work plan to the committee, and we phased our work plan into two phases. We had phase 1, which had three work streams. Work stream 1, which is the focus of today's presentation, was on consequential changes related to the STTR. including implications for treaty text and commentary and interaction with other provisions. Eamon is going to take us through that paper, which we'll also put to the floor for comments and for discussion. Then we also have work stream two and three. Work stream two focuses on the definitions within the model itself, including treaty definitions and broad cross-cutting relevance, and those identified as creating uncertainty. And most of these recommendations actually came from the previous committee, and they recommended that we need to look at them during this term. Then work stream three focuses on source rules. So what we are looking at within that work stream is concepts such as paid by, arising in, deemed to arise, the challenges that countries may be facing in that and whether they need any updates, especially in relation to passive income and cross-border services. So that is phase one of the work plan. And we'll just, we'll have a very quick overview of what those three work streams have done so far, and then we have phase two, of the work plan, which is actually going to start immediately after this session. We have three work streams within phase two. We have work stream four, which focuses on attribution of profits to permanent establishment, including issues under Article 7 and related commentary. And we had wide discussions on this yesterday. We had from the floor, and we also got to discuss the paper that was coming from the TP subcommittee, which was very helpful. We also have work stream five, which is dealing with the PE concept as a whole and the nexus rules, including consideration of non-physical presence and emerging business models. We also discussed this yesterday and Tuesday, I believe, and we also got some very helpful comments from the floor. And we also appreciated the paper that was presented by the DGE subcommittee. Work stream six basically discusses any other issues that don't fall squarely within the five work streams that we've already presented on. And yesterday we heard from Jorge and the subcommittee dealing with wealth taxation, and they indicated that they would be presenting a paper for consideration, and we are looking forward to discussing that within that work stream. Next slide, please. So across all the work streams of work stream of phase one, which was work stream one, two, and three, we've actually had quite a number of discussions since we had the 32nd session. For the subcommittee meeting itself, we've had two meetings. We had one in July and one in September, where we were discussing mainly the mandate and what is arising from the work streams that have already started their discussions. We are happy to report that each work stream now has a lead drafter and at least five participants across all work streams. Work has already started across the three work streams. For work streams 2 and 3, we held our preliminary discussions on 14 and 15 September. We also invited various stakeholders to contribute to those work streams, and we really appreciated the input that we got from from our observers, especially those who had raised these issues even from the previous committee. So we really appreciated the input and the insight that they had to present, and they've really been able to complement the discussions that we have in those work streams. We also had further discussions on 8th, 9th, and 16th of October, and specifically for work stream two and three. So this session is going to focus mainly on work stream one, which is the consequential changes to the STTR. In the next session, which is the 34th session, we'll then have papers from work stream 2 and work stream 3, as well as discussing work stream 1 further. For work stream 4 and 5, this actually came up quite loudly during the last session because there was some concern that we had pushed work stream 4 and 5 to the second phase. But we're happy to report that now that we have the transfer pricing and DGE reports, the second phase is actually starting immediately after this session. So we took your concerns and your comments into account and we actually updated the work plan that we had and we'd actually be very happy to hear about the change in timeline. So what we did is that we brought up the discussions under work stream four and five. So what will happen is after this session, we are going to actually have our first meetings next week. Work stream four and five will have the first meetings on 28th and 29th of October so that we can discuss the reports that were presented to us on Tuesday and yesterday to discuss the outcome of those reports and just have a way forward. So for work stream four and five, as I've said, the discussions will start next week. Then what we want to do is to ensure that there's going to be collaboration between the TP subcommittee and the DGE subcommittee so that we make sure that the work that is going to be done is going to be done in a coherent and consistent manner. Next slide, please. So now that we've discussed the background of our mandate and also what has been done across all work streams, the focus of today's meeting, as we've said, is the STTR paper. And we are happy to invite our lead drafter, Eamon, to present to us the paper on STTR. And then we'll come back to the floor for your comments. Over to you, Eamon.

Thank you very much, Wanjiru. And Good morning, all, and a particular welcome to our visiting students. The subject is the consequential changes related to one of the new articles that was introduced by the previous committee to the model tax convention. That is a subject-to-tax rule, and the consequences of that rule for other articles, the STTR appears in article 1, but the consequences of that new paragraph for other parts of article 1 and, indeed, other articles of the UN model. Next slide, please, Silvia. So, this morning's presentation will just start with an introduction to the subject to tax rule and then address the changes that have been made to commentary on articles 1, article 12AA and article 12C, which are principally made simply to acknowledge the existence of this new subject to tax rule, the paragraph 3 of article 1. And then we move on to the inclusion of an STTR derived low level of taxation approach to particular commentary options in the commentary on article 1. And then finally, we look at the limitation of benefits provisions in Article 29 and how the SDTR concept of low level of taxation can be integrated into the commentary on Article 29. So something I'd say at the outset is that all these consequential changes affect commentary rather than the actual article text of the model. Next slide, please, Sylvia, thank you. The headline is that the STTR is a really important addition to the UN model tax convention, and its purpose is to protect source State revenue. As you all know, tax treaties involve costly limitations on source State taxation in various articles, such as articles 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 1 12AA, 12B and 12C, there are limitations placed on the level of source State tax that can be applied to the payments concerned. The purpose of those limitations is to ensure that, overall, the burden of tax between the source State and the residence State is not excessive. The subject-to-tax rule then disapplies the source state limitation where the resident state charges low or no tax. So, if one of the contracting states is not charging the income concerned or charging it at a very low rate, then the need for the other state to limit its tax to prevent excessive tax or double taxation is clearly doesn't exist in the same way as if there was a sort of a standard rate of tax applied by both contracting states. So if the source state limitation was to apply in such circumstances, I think the possibility would be of excessive reductions in overall tax or indeed double non-taxation. And that would involve a wasteful loss of source state tax. And, you know, no countries can afford to sort of gratuitously give up tax where there's no need. And in particular, developing countries who are typically or often source, predominantly source states, need to be protected from unnecessary limitation of tax in such circumstances. So the SDTR relieves the source state from the treaty obligation to limit its taxation in circumstances where the statutory rate is very low or alternatively the effective rate of taxation is very low. So, there isn't specific reference in the paragraph three of article one to an effective rate of taxation, but by taking account of exemptions, exclusions, and reductions, uh, that affect the tax base, The paragraph takes account of very low or no actual tax paid, even where there is a high statutory rate, as a result of the application of those exemptions, exclusions or reductions. There is a focus on the statutory rate on the one hand and on the effective rate on the other, to determine whether, in fact, the resident state is charging the income concerned at a low level of taxation. The new paragraph 3 of article 1, which sets out the UN SDTR, is quite a distinctive provision in certain respects. It has a numerical or computational focus. It does not seek to describe uh, say, exemptions, exclusions, or reductions, it simply refers to them in gen- in generic terms, and then, uh, prescribes. a computation to be made. In the case of the statutory rate, it is simply to observe what the statutory rate is and see whether it is above or below a threshold rate that is set out in the paragraph. In terms of the effective rate, it prescribes a process whereby, ignoring any exemptions, exclusions or reductions, the threshold rate is applied to the income concerned, and if the actual tax paid is less than the amount of tax by that notional computation, then the effective rate is a low rate. So, it's a numerical or computational approach to determining whether there's a low level of taxation. The rule has a comprehensive, unrestricted scope, so it's not confined to specific types of income. It's a general rule that applies throughout the convention. And where the subject to tax rule has effect and the source state is not obliged to limit its tax, The source State can then apply its full domestic rate of tax. There is no restriction then on the application of tax by the source State. And where there's a low rate of tax in the residence country, that's an appropriate result because the intention is only to restrict the tax of charge by one state where it would lead to an excessive level of taxation or double taxation that would result in such an excessive burden. And finally, the STTR doesn't require that the parties, the payer and payee, be connected. it's sufficient that there's a low level of taxation in the resident state and in those circumstances the source state is not obliged to limit its tax on the income concerned. Next slide please, Silvia. So the consequential changes that I'm going to describe shortly take account of subparagraph C of paragraph 3 of article 1. So, paragraph 3 of article 1 is the SCTR provision. And paragraph A sets out the general rule that countries won't have to limit their tax where the receiving country has a low level of taxation. Paragraph B sets out how to compute what I've referred to as the effective rate of taxation. And paragraph C then provides certain exceptions. So countries might agree, say, that where a pension fund has a low level of taxation or no tax applied, that that won't be taken into account for the subject to tax rule, or other, say, collective investment vehicles or whatever. The countries, contracting states, when negotiating the treaty, can agree to create exceptions for particular reliefs or exemptions or reductions in tax so that they won't trigger the STTR. So next slide please, Sylvia. So now moving on to the second part, which is the acknowledgement or recognition of the STTR in other articles. This has already been done in many, many articles, but there were a few articles where it hadn't actually been, the SCTR hadn't been noted in the commentary. And I just again repeat, when I refer to articles, I'm actually referring to the commentary on articles, because all of these consequential changes are matters affecting the commentary on the articles concerned, rather than the actual article text. So, as you can see in the second bullet there, uh, the, uh, uh, existence and effect of the subject to tax rule has already been noted in articles, um, 10, 11, 12, 12B, and 13, um, so the articles that remained to be, uh… sort of addressed the commentaries to be addressed to complete the set where articles 12 AA in relation to fees for services and 12 C in respect of insurance premiums and all all that is said in each of those the commentary on each of those articles is that despite the income being received by the beneficial owner which would normally entitle that beneficial owner to the limitation of tax in the source state, if that beneficial owner is actually charged to a low level of taxation, they will not be automatically entitled to the limitation of tax by the source state. So that has been included in the commentary on article 12A and 12C. And in addition, In article 1, there is reference to the capital gains article, article 13, and the requirement that assets brought into a company that could have an effect on the percentage of assets represented by immovable property, and the requirement that the new assets be held for a period of 365 days, some text has been added to alert the reader to the fact that, if the resident's country has a low level of taxation, notwithstanding the 365-day requirement, whether that is present in the article text or not, the source State in respect of the disposal will not be obliged or may tax that disposal, notwithstanding whether the 365-day requirement is there or not. It is simply to alert the reader to the fallback provision, which is generally applicable, which, again, to repeat, is if there's very low tax in the residence country, then the source country will not be obliged to limit its tax. So next slide, please, Sylvia. So this is just the text in relation to the capital gains provision in capital gains commentary. which provide optional, provides an optional provision. It's a paragraph 159 of the commentary on Article 1 and the additional text drawing attention to the ever present, so to speak, protection of the subject to tax rule that's in the bold text at the bottom of the slide. So next slide, please, Sylvia. Again, this is just showing the text in article 12AA, drawing attention to the existence of paragraph 3 of article 1, that beneficial ownership of the income stream concerned does not automatically qualify the recipient for entitlement to limitation of tax in the source state. If in fact there's a low rate of tax in the residence state, then the source state will not be obliged to limit its tax. Next slide please, Sylvia. And again, similarly for Article 12C, the same idea, and I won't annoy you by repeating it once more. So next slide please, Sylvia. So now we move on to the Article 1 commentary in a different respect, and that's currently in the Article 1 commentary, there are paragraphs that set out optional provisions in relation to special tax regimes, and special tax regimes have a lengthy definition that spans a couple of pages, and various types of reliefs, reductions in tax are classified as special tax regimes. And of course, they're really close to the idea that's set out in the subject to tax rule in the definition of a low level of taxation. Because as I said at the start, the low level of taxation definition takes account, uh, both of both, um, a statutory or nominal low rate of tax. It also takes account of the possibility that there might be a low rate of tax through exemptions, exclusions or reductions that reduce the tax and the effective rate, notwithstanding that the nominal or statutory rate might be quite high. What the changes to the article 1 commentary do is to introduce, alongside and in parallel with long-standing provisions in relation to special tax regimes that are article specific, the concept of a low level of taxation. What the current commentary provides is that in individual articles, let's say article 11, the interest article, if the residence country, the receiving country, the payee country has a low level of taxation due to a special tax regime, as it's referred to in this commentary, the paying country, the source country, can withhold its limitation of tax, can withhold the treaty benefit concerned. And so, alongside the existing reference to special tax regimes, a new reference to a low level of taxation by reference to the concepts and the definition in paragraph three of article one has been included. And again, that's to promote awareness and use of the paragraph 3, Article 1, paragraph 3, subject to tax rule, the rule itself and indeed the components of that rule, in particular the definition of a low level of taxation. So next slide, please.

Sylvia.

The aspects of it that I would emphasise are that the changes are being made to what is already an option in commentary, and the changes will sit alongside the existing option without displacing it. Again, just to repeat that the new approach to those commentary options will not take account of whether the payer and payee are connected or not. In article 1, the existing special tax regime provisions are confined to situations that involve a connected payer and payee. That will not change; that will continue. Alongside the special tax regime optional provisions, there will now be included a further optional provision based on a low level of taxation, which will apply regardless of whether the payer and the payee are connected or not. The last committee decided that the UN subject-to-tax rule would protect source state tax revenues from unnecessary limitation, regardless of whether the payee in the residence state was connected to the payer or not. If the residence country was not charging a substantial.

Tax is so to say, and then the young girl, you know, you have to see. This is the whole apartment. You see, see, I pee the last apartment.

Let's.

That's the 4 changes.

This is the one that is about 29. It is limited 29 to provide the next year, you will 29. It's the longest. It's OECD and the United Nations范本当中最长的.

He got to sleep.

It is a long and difficult article, and it is complex and challenging. The points that I want to make about the proposed changes to the commentary are that They are simply changes to the commentary to that article. No changes have been made to the actual text of article 29. In the commentary to article 29, which includes the detailed version of the limitation of benefits article. In the commentary, there are further refinements or enhancements of that limitation of benefits article. They focus on whether tax-deductible payments made by residents who are seeking qualified person status, that is, persons who will qualify for the benefits of the treaty. The focus of the additional provisions in the commentary is on whether tax-deductible payments by persons seeking that qualified person status go to parties who are not resident in a contracting state or who are not resident in a state that is deemed to provide equivalent benefits under relevant treaties. And then thirdly, and this is the critical point here, whether those payments go to persons who benefit from special tax regimes. So, a factor that's take, an optional factor that is presented in the commentary to the limitation of benefits article is whether the base of prospective beneficiaries from a treaty is eroded by payments to low-tax jurisdictions. Now that we have a UN definition of low level of taxation, that is part of the UN model that the previous committee has decided to include in the UN model as the UN model's way of identifying a low level of taxation to protect source states. The commentary and these optional provisions in the commentary on Article 29 have been amended to refer to the low level of taxation concept as opposed to the special tax regimes concept. So next slide, please, Sylvia. So just a few quick points then, and I'm done. Again, to repeat the changes to Article 29, do not amend the text of the model article 29 in any respect. They relate exclusively to commentary on the limitation of benefits paragraphs, which are paragraphs 1 to 7 of article 29. They address only options in the commentary, and they address specific options and Those specific options refer to special tax regimes and to connected persons, insofar as the connected persons references have to do with special tax regimes. They insert, in place of the references to special tax regimes, references to a low level of taxation, which is, as I said, the new meaning and definition that has been included by the last committee in article 1, paragraph 3, in place of the references to low level of taxation in place of special tax regimes. Finally, where those changes have been made, where the references to low level of taxation has been introduced in place of special tax regime, there is repeated reference in each instance to the continued existence of the special tax regime option as part of the commentary to the 2021 model, which will continue to be fully available. So next slide, please, Sylvia. That is a recap of the various areas. I have just given the introduction to the SDTR, the acknowledgement of the SDTR in the few articles that are mentioned in article B, the changes made to the article 1 commentary and, finally, the changes made to the article 29 limitation of benefits commentary to substitute the low level of taxation concept for the special tax regime concept in the options in that commentary. So, thank you very much for your attention. I think there are final slides there as well. Thank you.

Thank you very much, Ayman, and thank you to Wanjiru for presenting the co-coordinator's report. I will now open the floor for comments and guidance from Committee members. comments from stakeholders. And I remind you that we are seeking input on the proposed amendments reflecting the STTR in the commentary to the model and also on the revised timeline that the co-coordinators have presented for the work streams. So the floor is open. Enrique, please go ahead.

Gracias, Trudy.

Thank you very much. It is a pleasure for me to see you here. We've missed you a lot, and we are happy that you're here with us. Well, I had the opportunity to follow this discussion in depth. This is a very complicated technical matter, and I'm convinced that this rule related to these taxes is crucial in comparison to other international organizations like the OECD that includes this rule, it is very weak, and the UN work is much more robust. And I also had the opportunity to discuss with Ajeon, and I'd like to congratulate him on this presentation. Very technical, comprehensive presentation, because this is not an easy issue, and I think he presented everything in a very clear way and I share everything that was explained by him and I agree with the point of view that is taken by the committee to approve the amendments to this rule. This is a cross-cutting issue. We have treaties to avoid double taxations. They're not looking at tax evasions, they're not looking at reducing taxations. They're simply aimed at redistributing these tax powers and the current document preserves the goals of the main treaties and the main goal being is to avoid multiple double taxations. Thank you.

Thank you, Enrique. Jorge, please.

Good morning. I'll speak Spanish. And good morning, Astrud. Very happy to see you here with us. I would like to thank for all the work that was carried out by everyone and especially Iman. And I also had the opportunity to participate in the discussions. And as Enrique just said, it was very clear from the presentation that this issue is very complex, very technical, and I believe that we have managed to move forward with what we already had on the basis of the previous mandate with the article or the clause that was introduced, how to adapt and include questions that were not yet reflected, and the fact that this objective is being achieved with all the work that was carried out by everyone, especially Iman. I think that is very good, and all the members of the committee were involved. I'd like to comment on two questions. I still have some doubts about these questions, and the first one relates to the possibility that The scope of application of article 1.2.3 may be limited by contracting states. It is clear that paragraphs 23 to 25, they're specific examples of possible exclusions. We're talking about revenues related to permanent establishments and other exclusions that are included in terms of pensions and other questions related to revenue. But it's important to reflect in the commentary that this doesn't cover all the possibilities as a result of negotiations. Contracting states may decide that there will be a more bilateral type of arrangement. And this should be referred to in a more explicit manner. And the second question that I wanted to highlight relates to Article 29 and the part of work related to this article. And what gives rise to certain doubts here are the references, for example, a subject to tax rule shares the same objectives and the rules may produce similar results. But I believe that any affirmation of this nature is not completely obvious, and I think that we should provide additional possibilities, other perhaps scenarios, as a result of this, and the reference to replace in this specific case in Article 29 the reference to special tax regimes with low level taxation is something that we should continue to analyze. If there is some possible result of this that may have unexpected consequences, I believe we should look into this somewhat more. Thank you.

Thank you, Jorge. Tan Santhiya, please. Oh, sorry. I saw another committee member here. Claudia, please.

Thank you, Trude. First, I want to congratulate the subcommittee for the work done, especially for IMO, for this paper of the work stream one. So on workshop one, I think that the proposed low level of taxation approach seems more objective and practical than focusing on special tax regimes. By looking at the actual level of taxation, it could provide greater certainty and be easier to apply across different types of income. Other aspect is the about the proposal to move away from the connected party requirements in a number of optional provision. I think that this approach appears consistent with the design of the STTR, which focus on the level of taxation rather than on relationship between payer and recipient. Aligning this provisions more closely with the STTR, the proposal could enhance coherence and consistency across the model. So, finally, on the -- about the work stream timeline, I support the proposed work plan, and in particular, the decision to defer substantive work on work stream 4 and 5. until the committee has considered the outputs of the DG and TP subcommittees. I think that given the implications of the ongoing discussions on economic nexus, non-physical presence and digitalized business models, this sequencing seems appropriate and will help ensure consistency across the committee's work. So I think that it's also highlights the importance of continued coordination among subcommittees to promote a coherent approach across related areas of work. So thank you.

Thank you, Claudia. Alexander, please.

Thank you, Madam Chair. It's forgiving to me a floor and it's good to see you. this morning. My intervention will be limited to questions which are mentioned in the co-coordination report on page 6, and they have some general nature. So I welcome this track before, and I thank the colleagues, especially Iman, for this work, for careful and diligent work on the supplementing the UN model and its commentaries with necessary provisions. We know that this Ctr. is a primary UN model response to a low level of taxation in the resident states. and was designed to protect source state taxation from the losses of revenue in circumstances of lower resident state taxation. Therefore, I believe work stream one input helps to ensure that all other provisions and articles of the UN model should recognize the SCTR and are aligned with it. I am in favor of inclusion of similar additions to the commentary on Article 1, 12 and 12, as we made in the 2025 Model Commentaries on Article 10, 11, and 12 and 12. To my view, the proposed wording provides necessary consistency and completeness for the next update of the Model. As for the Questions B and C, I think that definitely the addition to the model of the Article 1.3 of eCTD requires careful revision of the existing provisions of the model and the commentary to it as it is proposed in the subcommittee paper. I support the work on elaborating of the amendments or additions to the provisions on special tax regimes mentioned in the Article 1 commentaries, as it is in OECD model. Along with it, the new UN model should consider the consequences of the adoption of ECTR provisions for the commentary on Article 29. entitlement to the benefits, as Article 1 STTR and Article 29 Limited of Benefits have much in common. They seek to deny access to treaty benefits where recipients of payments can access special tax regimes or low levels of taxation. In conclusion, I am for consistency and coherence in the core concepts and approaches throughout the text of the model. and the commentary to the model. Thank you.

Thank you, Alexander. Angie, please.

Thank you, and of course, thank you to the co-coordinators and to IMEN for all the substantial work that has been done. And I think it's always positive to seek and to ensure coherence throughout the UN model, following the introduction, of course, of the STTR. At the same time, I am still a bit cautious about extending the logic of article 132 broadly throughout the commentary and other provisions of the model, replacing the references to special tax regime by the broader concept of lower level of taxation. may have consequences that also need further explanation because the concept of a low level of taxation is much wider and could capture situations that maybe were not originally intended by the provisions. Also, the proposal to remove connected party requirements in several places, I think it is justified within the SDTR, but extending this approach to other anti-abuse or the LOB provision is not as clearly warranted and requires further consideration and thoughts. So, but overall, also wanted to say that I'm a supporter of the sequenced approach. It appears sensible that work streams four and five proceed after the committee has considered, of course, the outputs of the TP subcommittee and the DGE committee on the P's and profit attribution. Thank you.

Thanks, Angie. Bashkar, please.

Thank you, co-chair, and thank you to the subcommittee for this very important work, because I think the STTR will prove to be a tool that will be of great use. I just wanted to make a few comments on the issue of this low level of taxation and the way it is recognized. Oftentimes, one would come across situations in real life when the, there is actually, it's not just a issue of low level of taxation. Sometimes in the other contracting state, there won't be a comprehensive tax system. And the, the, the, the taxation system of the other state will, will, will, will be subject only to certain classes of entities. So, you know, it, it, it doesn't squarely fall into the definition of low level of taxation or providing special exemption. So I think if it is possible, maybe the subcommittee can look at this situation also, that where the taxation in the other contracting state is not comprehensive, it is limited only to certain entities and a certain kind of tax regime has been brought in place. So because oftentimes a negotiator will come up against this problem. that the other state will say that his or her system doesn't fall squarely within the definition of the way low level of taxation is defined here. So that's just one suggestion. Thank you.

Thank you, Bhaskara. John, please.

Thank you, Mrs. Co-Chair. It's nice to see you. And thank you, Iman, for this very detailed and well thought out paper. I think it is important to adapt the commentary to the new STTR provision. And I agree with many of the amendments that are suggested. I have a couple of remarks regarding specific proposals. I don't know if you want to go through the pages of the paper. My first remark is for page 3. There's some change proposed for the next update of model, article 1, paragraph 159 of the commentary, where we're proposing an additional reference to the STTR regarding article 13. It seems to me from a purely, I mean, I fully agree with the substance of the article, but I'm just wondering, it's kind of strange to mention this in an article that focuses on the specific case of a 365-day period, which is the main focus of the article. And also just more in general and coherence of our commentary, it seems to me we could, of course, once you have a STTR rule, there's basically for -- there's many articles where you could say, well, wait a minute, there's a possibility if under certain conditions that here the STTR article applies. So it seems to me kind of strange to mention it here and not in the other article. So I'm not convinced that this addition is coherent with the rest of the treaty and really necessary, but that's a smaller point. My other point is regarding the proposal, the sector D for, in relation to article 29, and I think I joined some of what Jorge said and Angie and also maybe Abhikar. And this has not to do really with purely technical reasons that I can understand, but it's more maybe a philosophy of our commentary. It seems to me in the history of the UN and the OECD model, very, I think, rationally, when The provisions can be the same. That is really a plus for everybody. So the UN, we've often built off the OECD model because the OECD has done remarkable work on all these technical issues. And then there's cases where we vary because we have differences in philosophy in a certain way. But I think it should be our goal just to make the UN model as efficient and also easy to apply or easy to negotiate for us negotiators to try to as most as possible have similar sections and provisions. And when we defer from that when it's only really necessary. And here you kind of have the LOB block that is one thing. and you have the STTR block or article that is another approach. But at the end of the day, we're only talking about cases where you're having the alternative provision of LOB provision, which is not the most common situation. And it seems to me a very complicated to here, bring in, I understand that both, I understand the argument that both, both these provisions have the same worry about low level of taxation. But to bring in the STTR low level of taxation notions may be technically correct, but very complicated. And also, I think there might be some unintended consequences, as was mentioned before. And I would prefer to have the SDTR on one side and to keep Article 29 as it is, as the same as the OECD, or else it's going to, for a situation that's very rare, we're complicating it quite a bit. That would be my view. So, I would be in favor of not making the changes proposed in Section D. Thank you, John.

And now, we turn to Tanzania, who has been patiently waiting. Thank you. Go ahead.

Thank you, Madam Chair. Tanzania comments, thanks and job well done by the subcommittee. and welcomes the proposed amendments to the UN Model Commentary, particularly the recognition of the STR and Article 1, Para 3 as an important safeguard for source countries' taxing rights. However, we would like to suggest a few areas where additional practical guidance could make these provisions more effective for developing countries. First, Madam Chair, on the practical application of the STR, We suggest that the subcommittee provide guidance and worked examples on how source countries can verify the level of taxation in the residence state. For countries like Tanzania, obtaining information on foreign tax returns, preferential arrangements, exemptions, and effective tax computations can be challenging. Therefore, we propose a practical verification framework specifying the reasonable evidence that source countries may request to establish where the Ctr. applies. Second, Madam Chair, on source rules, particularly under articles 11, 12, 12A, 12B, and 12C, we suggest practical examples addressing situations where payments are structured through foreign entities. For instance, where a source country enterprise bears the economic cost of services, but payment is made by a foreign group company, where a digital platform collects payments offshore for services applied to customers in the source country, or where engineering services for a project located in the source country are paid through centralised procurement arrangements. These examples, Madam Chair, would help clarify how source rules should operate and prevent legitimate source country taxing rights from being undermined merely by the structure of payments arrangements. Third, on the meaning of low taxation, we propose that the subcommittee go beyond textual amendments and provide practical guidance on applying article 1, para 3, particularly in distinguishing low taxation arising from ordinary deductions, capital allowances, and carried forward losses from preferential treatment such as special exemptions, accelerated allowances, and contractual tax concessions. This should also address cases where domestic law limits loss deductions, but grants special incentives to selected companies. Such guidance is particularly important for developing countries with capital-intensive investments in mining, energy and infrastructure. Final chair on double taxation relief, we suggest guidance explaining how relief in the resident state should operate when the STRA permits the source state to exercise taxing rights. I thank you.

Thank you very much. John, your flag is up, but that's a legacy flag. Yes, thank you. Sierra Leone, please.

Thank you so much, Chair. I'm Janet from Sierra Leone. Firstly, I want to say congratulations to committee members for the great presentation and also course, how technical this work was, but yet seamlessly presented. I welcome this report from Sierra Leone. And this framework is also centered around the practical question we face as a country, in the sense that we will definitely consider how taxing rights of Sierra Leone can be protected, in the way that tax administration can implement tax laws and businesses can understand. Why this is of importance to us is the fact that this matters because we receive foreign investment and businesses make payments overseas for finance, services, and intellectual properties. The treaty treatment of those payments can affect the revenue Sierra Leone retains. So it is also important to note that we build on existing framework. The fact that Sierra Leone already has withholding tax obligation, transfer pricing regulations, and higher oil pricing guidance for transaction with associated persons. The proposed subject to tax rule will complement those measures. Transfer pricing examines whether the transaction is appropriately priced. The subject to tax rule addresses whether treaty restrictions should continue to apply when the relevant income is taxed below the agreed level abroad. So firstly, let me illustrate a practical relevance. Consider a business in Sierra Leone making a royalty payment on an overseas recipient. Our domestic law may provide for tax on that payment, while a treaty may limit the amount we can collect. If the income is then taxed at a low level abroad, an application subject to tax rule could allow Sierra Leone to retain taxing rights that the treaty would otherwise restrict. But that outcome also depend on the actual treaty and domestic law. Adopting wording in the UN tax model will not automatically change Sierra Leone's existing agreement. For us, therefore, the opportunity is greater protection of the revenue base. The challenge is making the protection work afterwards. I want to definitely highlight three practical needs. Firstly, evidence that can realistically be obtained. how the revenue administration in Sierra Leone established the tax treatment of a particular income abroad. A headline corporate tax rate may not answer this question where exemptions or reductions apply. The difficulty becomes greater when the payer and recipients are related afterwards. Secondly, businesses need clear answers on when to withhold, what evidence to submit, and how to obtain relief. or refunded. This matters because depending on the contract, additional withholding contracts increase the cost borne by Sierra Leone businesses. So, Mr. Chair, therefore, my recommendation is to support the direction of this work. accompanied by practical implementation guidance. That guidance should also include worked examples, reasonable evidence requirements, and clear explanation of how the rule interacts with existing treaty. For Sierra Leone, success will mean that the revenue administration can apply rule consistently, businesses can understand their obligations, and country can protect revenue while maintaining confidence in their tax system. I thank you.

Thank you very much. Nigeria, please.

Thank you, Madam Co-Chair. Good day, everyone. We want to appreciate Imon and the entire subcommittee for the job well done and for this presentation. We acknowledge that STR is a new rule. and a very technical rule at that matter. So many of us are still grappling with the challenges of understanding it and being able to implement it. So, and that explains some of the issues we are looking at and we are discussing and we are raising today. One thing, looking at the part A of the presentation of Imon, That presentation took care of some of the concern we are early identified in the discussion on the CRP of subcommittee on manual. So what we can only ask is that some of those points raised in that presentation. I think we benefit from it going into the manual so that, as we said, it's a technical area, we're able to understand how the rule works and how we can negotiate it effectively into the treaty. And also one part of it is the last part, I think the last sentence of the comment raised by my colleague from Tanzania, what happened when There is how to navigate the relief at the resident state when the source state has implemented the STR. So we benefit from guidance, especially in the manual, that can help us navigate those instances. One area we want to also flag is on replacement of special regime with low level of taxation. As we have said, it's a technical area, and we want to find out what would be the implication if the special regime does not does not result into low level of taxation because from what we have looked at in the paper, we think it's not all instances where special regime will result into low level of taxation. So in that instance, what happened? Because the moment we have replaced the low level of taxation, I mean the special regime with low level of taxation, then such instances, our understanding is that it will not be covered. So what do we do in such instances? We need clarity. Thank you.

Thank you. Yes, Enrique, you have an additional point. Thank you.

So sorry for asking the floor again, but I have a brief comment. because of the controversy that the low level of taxing is causing. I understand the concerns that Angie, Jorge, Vaskar, and Jan. I can understand their concerns, as well as what Kajinda has said from Nigeria. But I think that we need to find a way to keep the concept. And I understand the concerns, again. But I think that there's a very close relationship between the two concepts we're looking at here. I think that generally a preferential tax or special tax regime produces a low taxation, as a general rule that is. But there can be other cases as well when we design structures for some provisions in which it's not a special tax regime. because it requires voluntary act by the state to create a group of norms for different objectives. It could be to draw an investment. It could be to develop a certain region of their country, an area that's a bit behind technologically, for example. However, there can be other provisions. Because of evasion, tax evasion or other factors, they might lead to low levels of taxation. So I think it's very important to find a solution so that the concept is kept, obviously bearing in mind these non-foreseen unintended consequences or that can give rise to uncertainty. Thank you.

Thank you, Enrique. And now we have South Centre, please.

Thank you, Amaram, co-chair, and welcome. We would like to commend the work of the subcommittee on this important topic, especially important for developing countries. We'd like to align with the comments made by Mr. Bhaskar and by Tanzania, Sierra Leone, and Nigeria. And we'd like to begin by responding to the sequencing of the work, the proposed sequencing, and we welcome the change that work on work stream four and five will take place immediately after this session. And in fact, that is indeed the top priority for developing countries. So we strongly support this change. Coming to the CRP, we have just one comment, which is a question, really. If you look at the language on the changes to the commentary on dividends, interest, and to some extent, capital gains, the language says it would address the situation, the SGTR would address the situation caused by involving preferential tax treatment Whereas if you look at the language related to the Article 12 family, 12, 12A, 12B, et cetera, all of those articles, the language here says that the SGTR denies the limitation of tax in cases involving a low level of resident state taxation. So these are two different formulations for dividends, interest, and to some extent, capital gains. It says address the situation created by preferential tax treatment, whereas for the Article 12 family, it says denies limitation of tax in cases involving a low level of resident state taxation. so could the subcommittee clarify the rationale behind the differential language and the implications thank you.

Thank you um Enrique that is a legacy flag your flag is up I'm just looking around I see no further uh flags here Well, thank you all for your inputs. I think I will leave it first to the co-coordinators and perhaps Eamon to respond briefly to issues that have been raised.

Thank you very much, Madam Co-Chair. I'd be happy to try and respond. briefly, and I thank you very much for all your input. Enrique, thank you for your remarks. Jorge, I thank you in particular for your strong engagement with the subject in our discussions. Just very briefly, on the specific points that you raise, Jorge, I think that you mentioned that I was not entirely 100 per cent sure about the first point, but I understood that your concern was that the aspect of paragraph 3 of article 1, which is set out in sub paragraph C of that paragraph, which provides for exceptions for particular reliefs from the subject to tax rule, that more examples could be provided of those exceptions and there could be a greater stress on the fact that not all reliefs or incentives are going to be affected or captured by the STTR. And you're entirely correct in that, and I completely agree with your suggestion, and I think that is a matter primarily for the commentary and the manual on the subject to tax rule. I think the work presented this morning is primarily, sorry, separately about the consequences of that rule for other provisions, but I completely agree that there could be a greater emphasis in the commentary and the manual on those exceptions, and I think that can be addressed. I think that the second point that you raised, Jorge, is probably one that, you know, there's a considerable amount of sympathy for among colleagues, which is that the LOB provisions set out in paragraphs 1 to 7 of Article 29 and the STTR are quite different beasts, so to speak. and they should be kept quite separate. I will attempt to give reassurance, and perhaps we can discuss this further elsewhere, but the reassurance that I will attempt to give is that the changes to article 29 are extremely limited. To start with, They affect very specific options, and the very specific options they affect are options that focus on special tax regimes. So what is proposed is that instead of speaking of special tax regimes, we speak of a low level of taxation. So my essential point is that we are not affecting the architecture or the structure of Article 29. We are taking one very specific building block and we're saying, let's put a blue block instead of a red block. Apologies if that sounds too simplistic, but we are not affecting the article 29 limitation of benefits provisions per se. We are taking one very, very particular specific element of the commentary and optional element as presented in that commentary. And we're saying, now that the UN committee has set out a very specific form of subject to tax rule that was very intentionally and deliberately provided to protect source states, let's use that concept throughout the commentary on the model where appropriate. So that's the reassurance I would attempt to give Jorge, but I certainly hear the point that, you know, we need to perhaps consider further to what extent it's an appropriate adjustment to make to that commentary. But I would offer the view that it's a very, very limited adjustment that is very, very specific and it's replacing a very particular concept by a different form of that concept and it's not in any way seeking to affect the architecture or the structure of Article 29. I just welcome, Claudia, your remarks about the practicality or the computational approach. Alex, also, thank you. And If I can just try to pick out the questions here. I absolutely take your note of caution, Angie. I think you are probably quite close to Jorge's concerns. Again, I would hope that, with further discussion, we might be able to I might be able to offer the reassurance or the detailed explanations whereby these changes are very, very targeted. There is no question of some broad brush changing of the LOB provisions. I think, Vaskar, your comments were particularly interesting and the idea that countries may not have a comprehensive tax system. I do not anticipate that the STTR will immediately be included in all newly negotiated treaties. I hope that countries that need to protect their revenues as source States will take it up, but clearly any adoption of the STTR is a matter for negotiation between two contracting States. And if there are particular features of the taxation system of one state whereby it has an unusual configuration, so to speak, I absolutely accept, Vasco, that that should be taken into account in negotiations and the two contracting states will have to consider whether the STTR is an appropriate mechanism to include in their treaty. Uh, Gian, I, I would particularly, uh, like to thank your, uh, very strong engagement with the discussion, uh, throughout, uh, this process and, um, Just coming to the particular points that you mentioned, you mentioned capital gains and why you refer to STTR in the capital gains paragraph in article 1. There are many, many other instances where we could have drawn attention to the STTR, and I think it's a valid point. But just as some other points were raised in our discussions in our subcommittee And adjustments were made, say, for instance, to refer to sub-paragraph C of paragraph 3 of article 1, the exceptions for special for incentives, et cetera, allowed by that paragraph, and, for instance, suggestions that we make continued reference to the special tax regimes in article 29. There was a suggestion that we make a reference to the STTR in relation to that particular paragraph. I take your point that it may be perhaps not entirely in line with the absence of a reference in other instances, but equally, I think it does no harm and it is pointing out a factual position whereby, if there's a low residence state rate, then that will also be an answer to the difficulty that a country might, the concern a country might have that it would be inappropriately giving up its taxing rights on shares deriving their value from immovable property resident in the country concerned. On article 29, I think that it is largely what we have discussed previously. The discussion of the philosophy of the commentary in respect of the UN and the OECD is potentially very interesting. I think that that would be Uh, good, but, uh, I'm also conscious of time limitations, so I, I just note the, note the, um, points made there, uh, I do believe that the committee, the last committee was very, very intentional and deliberate in providing a particular form of STTR. I prefer, when I am at the UN committee, to speak about what we as a committee do here, rather than engage too much in a reflexive comparison with other developments elsewhere. I know that what we did and what the last committee did was done very intentionally, and I think that that is should be reflected in the commentary where appropriate. Okay, I'm told I'm running out of time here. So again, thanks to Tanzania. I think most of the points that I heard from Tanzania are entirely valid, but equally I would say they are points in relation to the article, sorry, the STTR itself and the commentary and the manual on the STTR. And again, this morning, the work is very much limited to consequential changes as well in the model, but I entirely take the points that you make and their relevance to understanding of the SCTR as in its own right. And again, I just welcome the comments from Sierra Leone, in particular the example in relation to royalties. I take the point made by Kehinde from Nigeria, the very specific point in relation to perhaps a special tax regime might not be covered by the low level of taxation. I did certainly look very carefully through the definition of the special tax regime. I think it's a very interesting point and I think it does merit further examination. So I think then the South Centre, I again, I, sorry, the last question, if I can just briefly look at what was involved here. Yes, prefer, different use of language. I think that preferential tax treatment would generally lead to a low level of taxation, but certainly we can have a look at ensuring consistent language, but I think certainly the thrust and the substance and the intention is is the same there. So, apologies for taking a little bit of time in trying to respond there. And, John, did you...

Thank you, Iman, for the presentation and also for the responses to the questions that have been raised. And thank you all for your comments and questions as well. We would also welcome written comments so that we can be able to comprehensively analyse those comments. On the second issue, we also welcome and appreciate the support that we've received in relation to update of the timeline. And it actually aligns with the comments that we received in the 32nd session, that workstream four and five are very critical and we need to start those discussions as early as possible. And now that the committee has considered the reports from transfer pricing and the digital economy subcommittees, those discussions will actually commence next week. So, thank you for supporting that timeline update. On the implementation guidance raised by Tanzania and Nigeria, we understand and we agree that this guidance would be very beneficial because this is a very technical area that would really need to be explained well to developing countries for implementation. And we see that this could be done actually throughout this work. This could be done through capacity development, and we can discuss that also with the team in charge of that. We could also do that through coordination with the UN manual on how this could be included in the manual to provide guidance, and also dissemination through capacity activities on a virtual basis. So we appreciate that comment, and we will look into how that can be done. Thank you so much, co-chair. We can hand back over.

Thank you very much to Vangero, Antoine and Eamon. There was strong support for ensuring consistency and reflecting the STTR where appropriate throughout the commentary. However, also some points for further reflection and we look forward to to those discussions and the next iteration of this paper on the STTR. And also, there was strong support for the revised sequencing of the workstream. So thank you all for your input and engagement. Written inputs are welcome by 26th of November, but I also see John's flag, so go ahead.

Thank you, Imen, for the very detailed responses. Just regarding the timeline, it wasn't clear, I mean, yesterday we were talking about Article 7 being modified and maybe the UN model group doing that or not. was wondering because there would be a big impact on the timeline presented, I was just wondering when we're going to discuss this and decide this, is that still going to be in this session or is it further down the road? Is it the committee who decides that? Are we going to discuss this in the subcommittees? I was just wondering what the way forward was regarding that issue.

Thank you, John. I think the division of work on cross-cutting issues we can discuss further in the closed meeting, but of course, the work stream four is very well situated to take up what we discussed in the transfer pricing subcommittee yesterday. So, thank you. With that, I think we'll break for coffee for 15 minutes. Enjoy.

Test on glade, test on glade, test on glade, test on glade.

If everyone could take their seats, please. We will move on to item 3E, tax administration and artificial intelligence, and this is CRP 16. So we will have the co-coordinators present the progress of the work and the detailed outline for practical guidance to implement artificial intelligence for tax administrations. So, Geneva and Alexander, please go ahead.

Thank you, co-chair, and welcome. Great to see you. So my co-chair and I, Alexander, will actually share the podium this morning. I will start the presentation and then he will come in in the middle to complete our presentation as to the progress of what we have done so far since the committee subcommittee was approved, established, and then the work plan was approved by the committee. So quickly, we will I'll walk you through the background, the proposed outline of our practical guide, and then, of course, the action that we will request of the committee as to how to progress on the next steps going forward. Next slide, please. As a background and mandate, the sub-committee is mandated to prepare a draft practical guide to implementing artificial intelligence for tax administrations, giving guidance on matters such as fraud detection, risk assessment and mitigation, taxpayer service, guidance frameworks and safeguards, including data and gender biases, and promoting integrity as well as data confidentiality. Our subcommittee is made-up of 28 participants from diverse backgrounds, including committee members as well, academia, tax administration, the private sector, and international organizations such as including OECD, CIAT, and ATEF. As far as milestones are concerned, the subcommittee was established at the 31st session a year, exactly a year ago this month as we sit here. And then in New York, we presented the work plan and where the work plan was approved, a practical guide, chapters and things like that. So we are here now today to present to you the progress that we have made so far. And then in the 34th and 35th sessions, we expected to present our first and second drafts. Next slide, please. To implement the work of the subcommittees as approved by the committee, we grouped the subcommittee participants into either draft or review, one or more of the chapters of the practical guide that we will later explain and walk you through. and then what we have done, the subcommittee actually has met three times so far. We met on the 17th of April and also on the 16th of June and then the 31st of August. But of course in between the groups met, the groups that we set up met to just make sure that the work progressed. Next slide, please. So when it comes to the practical guides, we have five chapters. So chapter one is introduction, of course, which is the window to the practical guide. Chapter two will talk about the key enablers. Most important of them is, of course, the technology. It is important that you have the right infrastructure. to enable the implementation of AI. Of course, the third chapter talks about AI use cases, which was important that we identify and make sure that we have examples of what is happening out there. Safeguards, chapter four will focus on safeguards. And then on chapter five will be the conclusion that will bring all the guidelines, practical guidelines together. Next slide, please. So chapter one, as I mentioned, is the introduction. And basically what that chapter we propose will do will actually be the window to the guide by outlining its purpose, intended audience, and then scope. The chapter situates the topic within the broader digitalization trends in tax administration, while navigating practical constraints and opportunities, particularly for developing countries. It summarizes the core building blocks needed for effective artificial intelligence implementation, such as a robust infrastructure, as I mentioned earlier, data readiness, staff capacity and skill set, legal and regulatory framework. It then explores what AI means in a tax administration setting, distinguishing it from related digital and analytical tools, which is very, very important. The chapter then will close by setting out the key drivers for AI adaptation by tax administrations that will be further explored in chapter three. With that said, I will now invite my co-chair, Alexander, to take it from here. Thank you.

Thank you, Bangura. I do apologize for my voice again because three days raining have made me a little detail. But I do hope just I will manage that difficult situation and will present you next slides. So having taken-- yes, thank you. Having taken the mandate given by the Committee and preliminary considerations with respect to the purpose, scope and targeted audience of the practical guide, as a basis for further discussion on what should be done to elaborate on the implementation of AI for the tax administration, the subcommittee would propose to start the discussion with the so-called building blocks, key enablers for responsible AI adoption. Chapter 2 will describe these key enablers that need to be in place for the safe and effective adoption of AI. Without digging into details, we will definitely do that when we start the guide drafting, the subcommittee identified following enablers. The first one is technology and infrastructure which needed for implementation. The second one is the data. The section in the chapter two will give a user of the guide to know about minimum data foundations for AI to function reliable and efficient. The above mentioned enablers we consider as a ground to build AI infrastructure for the tax administrations. The third enable is a human resource considerations. It is kind of, we believe, a lot to give an opportunity for AI to operate. The fourth, fifth, and sixth enablers are legal considerations, governance framework, and last but not least, phase implementation approach correspondingly. Next slide, please. Chapter 3 will deal with the issues related to the practical ways AI can support core tax administration functions. The subsections of this chapter will provide practical descriptions And in country examples demonstrate how tax administrations can use AI to effectively manage taxpayers registration, improve services, render it for taxpayers, identify areas for the tax audit, better prepare for tax disputes, and finally, improve revenue forecasting and making other analysis. Next slide, please. Chapter 4 will provide details on main risks that may arise when AI is deployed by tax administrations and will present practical ways to prevent this risk or mitigate them if there are no possibility to avoid these risks. In particular, the drafting team has identified risk associated with the privacy issues, and data protection, discrimination risks, errors and system failure. A specific subsection will provide for overview of how tax administrations could detect, correct and remediate errors in AI system. Again, the last but not least, a special subsection of the chapter four will provide overview of how tax administrations could deal with the taxpayers' rights for instance, in the areas where it concerns confidentiality matters. Next slide, please. The final chapter five provides for some thoughts of how to operate AI and how to use AI in our fast-changing environment. In conclusion of our presentation, I would like once again to note that all chapters will include practical examples and case studies from countries at different level of AI adoption. In this respect, the subcommittee would highly appreciate if committee members, representatives of the countries, business society, and other observers would share their country examples, which allows us to make the guide much practically possible. Next slide, please. Now I would like to thank you for your attention and invite to provide comments and guidance on the detailed outline of the practical guide presented for consideration during this section in the UN Tax Committee. Thank you very much. Maybe, Bangor, you would like to add something?

Thank you very much, Alexander, for that cooperation and coordination. So I will now like to pass it over back to you. co-chair, for the next steps. Thank you.

Thank you very much for the presentation of this detailed outline. I open the floor for comments and guidance from Committee members and observers, stakeholders. Angela, please.

Thank you, co-chair, and thank you to our co-coordinators As a subcommittee member, I would like to say that it has been a privilege to actually witness the passion from a very wide group of stakeholders in drafting this piece of work. They have offered expertise and experience, and there's just so much that we want to put into one guide. And so as a subcommittee member, clearly I support the work that has been laid out. The coverage of the guide is comprehensive. We cover the enablers, the use cases and the safeguards, which is a very important part of using AI. I just have two points, points that I've also highlighted in our subcommittee meetings. Just as Alex has mentioned, I support the focus on examples and case studies. so that we can make sure that the guide is practical and usable across all levels. It is hard to find a one-size-fits-all approach, but as a tax administration, we like to learn from others, and you learn from seeing other tax administrations or other organisations that have implemented AIs, even at different levels of AI readiness and maturity, whichever level you are at, is still able to access. And I think that that would build confidence and will be useful for developing countries. My second point, and it's a related point, is on including examples on the use of generative AI. The reason I say this is because traditional AI use cases are greatly valued in tax administrations, but these use cases tend to set a rather high bar for entry in terms of implementation considerations. So with the growing availability of Gen. AI tools in the market, this could help developing countries and have a easier access and a practical pathway to implementing AI use cases without having to develop AI solutions from scratch. So personally, in my personal experience and in our organisation, GenAI use cases has really enabled staff across all levels to start using AI solutions to improve productivity. So these are my two suggestions and I'll continue to work hard together with the subcommittee. Thank you.

Thank you, Angela. Enrique, please.

Thank you, Trudy. First off, this is a tremendous universe technology. There are a few -- there are so many questions here. I commend the entire team and subcommittee, each and every one of you, for this work. I think that technology has exponential progress or advances, and it opens up a wide array of enormous possibilities. I think also that it's very important to continue to follow up on changes in AI, its evolution further. It's a very short time period. In a short time period, it makes a great deal of progress. So it is in this context that I would like to explore the possibility, and I put this on the table, specifically a neurosymbolic AI. I see that the document addresses generative AIs, those that have the capacity to learn from massive data sets, deep data, Gemini, ChatGPT, Claude, etc. But AI, this neurosymbolic AI, further has beyond the ability to learn from deep data. It also has the capacity of undertaking reasoning called symbolic. So this technology already transcends simple probabilistic calculation. It takes on the ability to reason with logic and even to draw conclusions that are solidly founded or grounded, I think that that's one step beyond generative AI. So why am I putting this topic on the table? I think it's important that tax administrations begin to take a step towards this evolution in AI because the private sector the big consultancies, they're already providing services using neurosymbolic AI tools, consumer goods sectors, financial sectors, many of them. So in practice, the private sector is already using such AI. Now, with that, I'm not seeking to say that we not delve deeper into generative AI, which I think the guide mentions and correctly so, or that will be eliminating human opinion or involvement. That won't happen. But I think it's worthwhile, if you agree, that we could enrich the document with this other type of AI. which again is already being used by the private sector. It would be a good idea, advisable for the tax administrations to know of this and to use it as a useful tool for meeting its different mandates. And that the guide mentioned these tax collection implications and processes, et cetera. That would be all. I appreciate it.

Thank you. Shanice, please.

Thank you so much, Chair. As part of the committee, I definitely do want to commend the work, the excellent work that is ongoing within the committee. And I do want to make the point that as a committee, it is noted that AI adoption is not about technological expertise necessarily. It's not simply a technological exercise. The reality is for different developing as they seek to incorporate AI into their processes at whatever level. It is changing the roles within the organizations, it's changing professional identities, it is changing career paths and the responsibilities of tax officials. So it's very important that tax administrations make a deliberate human resource allocation change management system to ensure that the AI itself is being able to be utilized by the persons within the organization. And it makes sure that the organization has to develop systems and processes so that tasks which can be automated are automated and those which can be enhanced by AI are enhanced by AI, whilst also remaining human-led. Because it is important to retain the fact that human-based decisions provide fairness for taxpayers, it provides transparency for taxpayers, it provides systems of appeal, It ensures that we are able to manage tax administration and effectively, whilst also giving taxpayers a route and a mechanism for appeal and making sure that the process is fair. And so it ensures that we consider that investment in AI and data science skills and change management skills and ethical governance and development of tax expertise within the realm of AI has to be focused on these factors. It has to be focused on the human led, the human decision making factors while utilizing the processes that AI provides. And I think that's one of the key enablers that is discussed in the in the subcommittee, the human element and the deployment of human resources at various levels in different developing countries and in different developing contexts are very important for us to look at on a continuum from tax administrations who do not have as many resources all the way up to developed tax administrations and acknowledge that there are different skills necessary, whether it's to be sourced internally or externally from the tax administration is a decision that has to be taken. But it is a decision that is worth appreciating before we can start to deploy AI in tax administration. So that is a key area for the committee and something that we have discussed at length. And it is something that is going to be a central theme and thread throughout the guidelines and just something that I want the committee to be aware of, the fact that we have given a lot of thought to this matter. That's my intervention. And again, excellent work by the committee. Thank you.

Thank you. Darlington, please.

Thank you, co-chair. Thank you. Let me join Shanice to say such excellent work done by the subcommittee. The two co-coordinators, thank you. Being part of the subcommittee itself and being a tax administrator for a couple of years, I think the subcommittee on AI and tax administration is so vital to the work of especially developing countries mobilizing more revenue in the individual countries for national development. Having said that, just one quick intervention, especially on section three, particular reference to section 3.5. I think it discusses the constructive manner in which AI can be used by taxpayers in meeting the obligations, i.e. filing, registration, payment of the taxes. I think and I would suggest that it goes a step further as well to discuss the misuse of AI by taxpayers in in filing um false documents in um using AI to deliberately file false documentations as well I I know the the the section talks like I said talks about the constructive manner in which AI can be used by taxpayers um but I would suggest I also go further um recent a recent report I think I read um being discussed by the HMRC discusses how taxpayers are beginning to use AI to falsely declare their tax obligations, i.e. whether they are filing, whether they are paying, or whether they are registering companies. So I would also like further discussion to be embedded in section three as it is right now. My last comment has to do with probably strengthening section four that has to do with, um, um, cybersecurity. I, I know, um, recent development in AI, open AI, that, um, cybersecurity is, is, is, is, is trending in, um, the AI space right now. So I will also paint that section three and section four also being strengthened. that tax administration can be made aware of the threat of cybersecurity in the use in the use of AI in tax administration. But let me say once again, thank you to the the co-coordinators and and I look forward to contributing to the work of the subcommittee in the days and and months to come. Thank you, Madam Co-Chair.

Thank you. Uh Niri, please.

Thank you, Chair. First, I join with the colleagues regarding the congratulations to our co-chair, co-coordinator, and the subcommittee member on the outcomes of the detailed outline. And I totally found it very detailed. And I support it because of the-- AI currently in my workplace, everywhere people talking from the way to the gate to the office, everywhere people are talking about AI and they are now learning using AI, especially on the tax administration. So I totally support for this. I have one clarification, especially I much appreciate on the chapter three, regarding the AI use case. I see the flow very comprehensive, you starting from the clarification, registration until the dispute settlement. But one thing that I feel myself missing from it's about the investigation procedure. So I doubt myself whether is AI applicable for the investigation procedure and as everyone know that investigation procedure for tax issues is very complicated and very confidential because some cases are moved up to the court. So is it possible here we have a practical guidance on the investigation in this paper? And thank you again and I totally support the work done by the subcommittee. Thank you.

Thank you. Angie, please.

For the excellent progress made on this very timely and very ambitious project. I strongly welcome the proposed outline. I appreciate the balanced approach that recognizes both the opportunities of AI, but of course also, which is very important, the safeguards, as Shanice also mentioned. I would also want to encourage the Subcommittee to further explore the practical examples and case studies, especially on how tax administrations can remain accountable when AI systems are used to support audit selection and dispute management. We believe that transparency and explainability is key, I believe, that it's key for maintaining taxpayer trust. because AI can strengthen tax administrations, but trust must be in the center of the way we use this framework. It can support all our decisions, but the end responsibility, I believe, must still remain human. Thank you.

Thank you. Hector, please.

Thank you, co-chair.

Just a really brief comment. I commend the subcommittee and the drafters. I do think this is a very important topic. Just maybe picking up on a couple of the points made about, you know, I phrase it as humans make decisions and in tax administrations, it's really important that humans are making decisions. And AI, the uptake of AI is accelerating significantly and I think one of the challenges that is becoming increasingly evident for tax administrations is that in the tax administration a human needs to make the decision but there is a lot of AI generated information coming at tax administrators and so there's just a capacity issue that I think is really important to call out in the work. Thank you.

Thank you. Bhaskar, please.

Thank you, co-chair, and I congratulate the subcommittee for this work that is being done. And yes, I'm sure that any AI-driven tool that can be used will increase efficiency. But just a couple of quick points, especially in the context of developing countries and some of the experiences that we've had. Probably there is a case for deciding what is the right time to introduce any AI based automation because sometimes if it is too early, then some of the decision making can be led astray. And this is this is this is the experience that we have because especially in developing countries where if there are a large number of taxpayers, different kinds of taxpayers, there is this balance that has to be created between development and disruption because sometimes if any technological based change can distort the kind of decisions that are taken, especially in the matters of tax assessments, et cetera. So this is another point. The other point that I would like to make that any AI based system, especially for risk management, it depends largely on the kind of data capture that you have. And this is an experience that we've had, especially We have used some sort of tools for picking up cases for transfer pricing audit. But it succeeded only when we were able to actually prescribe the right kind of data that should be made available in a basic form that was that you know for for reporting an international transaction etc so it's only when that that level of data capture is ensured that an AI based tool can effectively operate on that to pick up cases for audit effectively so these are a couple of things a couple of suggestions which if the subcommittee can keep in mind maybe the guidance will be a little more effective especially for developing countries thank you.

Thank you. John, please.

Thank you, Madam Co-Chair. Thank you very much for this important work. It's indeed such a revolutionary, I must say, subject. There's the contrast between the mind-boggling possibilities of AI that are fantastic and also the mind-boggling risks that go with this. And I just want to say that I approve this work, and I thank all the committee members for their dedication to this work. I also join previous comments on the, of course, the importance of human decisions. I was wondering maybe if we would want to add a building block in creating an AI tool that could fairly allocate profit attribution to SEPs, but maybe that's a little ambitious.

It would be helpful, though. Sierra Leone, please.

Thank you, Chair. Firstly, I would like to commend the work of the committee members and also the co-coordinators. considering the fact that the presentation was short, but quite explicit, it shows that the committee members had put tremendous work to come up with a clear cut out summary. Firstly, Sierra Leone welcomes the report and proposed practical guide. However, the main concern will be how an administration with limited resources can adopt AI successfully and sustain it over time. The implementation of AI in tax administration also mean advancing digital transformation. As a country, definitely we have made series of intentional effort to implement digital reforms through the integrated tax administration system, the world amongst others. We have also giving foundation to, these are some of the foundations we have built on. However, we encourage the guide to distinguish between having digital system and being ready to use AI effectively. I would like to highlight four practical guides from Sierra Leone's perspective, which I also know this is a shared experience we have with other developing countries. Firstly, infrastructural and reliable energy through electricity supply. Consider firstly a business whose electrical cash register system stopped transmitting because of a power outage or poor connectivity. An analytical tool might flag the missing information as a compliance risk. the officer must first establish what caused the interruption. That is why I think that the guide should provide practical options for back-up power, affordable connectivity maintenance and continuity of service. It should also recognise that some administrations still need investment in those foundations. The second issue is skill development and meaningful knowledge transfer. So, firstly, we also have to understand that most of those tools are built, maintained for a certain period of time, and then transferred to the host country. When external support ends, can our officers operate the system, questions and resolve problems? This is another concern. What should be measure of successful technical assistance? We would welcome guidance that makes joint implementation, training of trainers, practical documentation and continued monitoring part of project delivery. Capacity building should reach tax officers, ICT specialists, legal team managers and taxpayer services staff. Third, a phased approach with affordable starting points. From Sierra Leone perspective, an initial application could help officers find approved tax guidance and answer taxpayers' inquiries more consistently. This would also allow us to build experience within controlled setting. More complex applications such as identifying discrepancies across returns, customs information, ECR records could follow when data quality and staff capacity permits. Each pilot should answer a clear question. Does this tool improve the services or solve the problems at a cost we can sustain? That's another concern. The assessment must also include electricity connectivity, maintain continued staff development. Finally, I think also AI generated indicators should support professional judgment, decision affecting liabilities or enforcement must maintain grounds in law, verified information and meaningful human review. We must also preserve accessible assistance of small businesses and taxpayers with limited digital access. That is also another concern. Furthermore, co-chair, coordinators and committee members, these priorities lead to specific request. the guide should offer different implementation pathways for different levels of readiness. A practical readiness checklist example for administrations with limited resources, realistic cost guidance, and measurable knowledge transfer request for successful adoption will mean that our officers can use system confidently our institutions can maintain it and the taxpayers experience a more reliable and fair services. I thank you.

Thank you very much. Singapore, please. No, no, first I have Gidrey. Sorry, I saw her flag.

I will be I will be very quick. I just wanted to thank really the co-coordinators for already very big work which was done in the last year in organizing meeting and getting together all those stakeholders who gave also technical inputs which were, you know, for tax persons not always easy to understand, but the inputs were very clearly given and they learned myself a lot. So I Just wanted to maybe continue what was raised by stakeholders and some of my colleagues about the cost as well. And I know the report, the guidance will cover this, but as for me, this question is very important. I wanted to I think it's really important to stress it in the plenary session. So, AI is very expensive and there is this hard dilemma for a lot of countries, whether we develop our own AI in tax administration or we use something foreign. And with using something foreign, there is this information security question, which was highlighted briefly in the document and then we discussed it during the meetings. But I just wanted one more time to highlight that there are legislations in certain countries that allow access of their IE companies to data and prompts which are even held in service in other jurisdictions and those important questions should be, information security questions should be really put in the guidance as in one of the first like priority places in my opinion. But thank you very much for all the work, and I will continue to follow the Subcommittee with a lot of interest. Thank you.

Thank you. And now we move to Singapore. Go ahead, please.

Thank you, Chair. Singapore thanks the Subcommittee for the hard work on this very important topic, and reiterates our support for the Practical Guide. We have two comments today for the subcommittee's consideration. So first, we think that humans remain as the most important part of the workforce. The guide rightly recognises in section 2.3 that human resources are a key enabler for responsible AI adoption, and that successful adoption depends on balancing external AI specialists with reskilling the existing workforce. So it is important for tax administrations to support the workforce in adapting to new technologies. The guide stresses the role of AI as a tool to enhance operational efficiency while preserving human oversight, and highlights the importance of future-proofing and continuous learning. So we suggest that the guide could also include how tax administrations can progressively redesign roles, preserve critical tax expertise and institutional knowledge, while creating realistic pathways for staff to move towards higher value work that complements AI usage. Second, as AI systems become more autonomous, no single safeguard is enough. We hope that the guide continues to take into account the different needs and stages of AI adoption and provide practical and proportionate guidance to help tax administrations build trust, manage AI risks, and adopt AI responsibly. We hope that these two points can be considered as the drafting progresses. Thank you.

Thank you. And now we have Claudia, please.

Thank you, Madam Chair. I just want to congratulate the work done by the subcommittee. This is highly relevant and timely topic as AI has the potential to enhance efficiency, improve compliance, and help tax administration make better use of often scarce human resource. From Brazil's experience, we can already observe the benefits that these technologies can bring when deployed carefully. Our tax administration has begun using AI-based tools and has demonstrated that it's possible to achieve significant gains in effectiveness and service delivery while maintaining appropriate safeguards to address potential risk. So in this regard, I particularly welcome the work taken by the subcommittee. It's essential that the adoption of AI be accompanied by robust governance frameworks and appropriate human oversight, especially in critical stage of decision making process. Brazil experience shows that AI can be powerful instrument to support tax administration, provide that human judgment remains central to decision and that due attention is given to transparency, accountability and risk mitigation. So congratulations. Thank you.

Thank you. I do not see any further flags. However, I know that we have some observers that may not flag. Ah, yes. And who is? Go ahead and please present yourself. Angela. Yeah, it's, it's, now I can see you. Go ahead.

Thank you very much, Chair. Um, I wasn't sure if I was raising my flag right or not, but we're here. Just to congratulate the committee. This is very important work. In the previous membership, as you might recall, we wrote on and researched on digitalization, tax administration and digitalization for revenue authorities. And the question came on how far we would go, whether we would go into the AI topic or not, and we safely postponed. So now it's good to see it being picked up again, and especially as timely as it is, right? And one of the concerns we had when doing this work was to see what is the role of UN guidance on a topic such as digitalization for governments or a topic such as AI. And if I dare to give my opinion, I think the role of these documents is to be as targeted, as clear, and as practical as possible, because tax administrations and coming from a tax administration also, I know the difficulties in and the overload in some of the materials that we have to go through to perform the work. And so as targeted and as clear as this guidance can be and providing practical pathways, I think this is how value can be added for tax administrations. That is one point. The other point is, and this connects to a bit of the discussions that we have been having during the week, regarding ambiguity and mentioning all possible alternatives. I am of the view that the guidance produced by this committee should be clear and directed, not so much not so broad and presenting all possible avenues. I think the committee would pick a direction and then stick with it and guide tax administrations with it. That's how we can add value. The second point concern that we had at the time and that I think it could also be relevant for this work is Concerns about human rights and constitutional taxpayer guarantees. This is, I think, the main role of this kind of guidance, because, of course, there's much written on and researched on AI and AI tools. taxpayer tax administrations can go to those those sources for that. I think when a tax administration or an official is looking at UN tax committee guidance on AI, they would probably be asking themselves the question of what what do we have to what safeguards do we need to have? What should we be concerned about in in applying these tools? Because these tools will be applied are being applied. And so human rights and taxpayer guarantees, I think, are a big concern, should be outlined very clearly. In the era of data being the biggest, one of the biggest commodities and resources for companies operating online, how is taxpayer data going to be protected in the face of large learning models and AI that is self-generative, right? That's one of the big concerns, is that AI tools can, now they can self-generate and self-improve. So how do we as governments protect our taxpayer data in the face of these risks? So it would be ideal, I think, as a suggestion to point out these risks. We don't necessarily have to provide all the solutions, but to flag out some of these concerns. Another concern that is also widely discussed are the environmental concerns and risks of adopting large AI tools by governments. And a third one would be, uh, the financial cost, which is the big elephant in the room, everyone, I think all tax administrations. want to have the most developed technologies. We want to be more efficient. But what is the cost of that for the different levels of development, different economies, right, that are applying to this? And again, this doesn't have to provide the solution, but point out, as I think a previous speaker said, the different levels of implementation, right? It has to be said, I think, when a tool that comes from the UN Tax Committee doesn't mention the risks, I think it really, it doesn't fulfill its full purpose. So, yes, so mentioning these concerns as clearly and as loudly as possible, I think would be a great value addition. And yes, and I think that's all I wanted to contribute to congratulating the work and then very excited to see what the tool will look like at the end. Thank you.

Thank you. And it's very nice to see a former member here as a stakeholder. So thank you very much. I will repeat what I said. If there are anyone who does not have a flag but would like to make a comment, you can raise your hand. I don't see anyone. Well, thank you all for your input. I think there are broadly... The themes are balancing the risk and opportunities that AI presents for tax administrations. also giving guidance adapted to different levels of capacities in developing countries, and also maintaining the human decision making and having AI as a tool rather than decision making. I think you have had some very good input for issues to be elaborated in the guidance. And Geneva and Alexander, if you would like to make some further comments before we close this item, please go ahead.

Thank you very much, Co-Chair. I just want to say thank you to each and every one of you for your contributions. I mean, as Co-Chair has just highlighted, for me, the takeaway is as tax administration, AI is important and we need to keep up with what is happening in the AI world. But also it is important that the human influence is important, right? As much as we want to keep ahead of you know, what is happening in the world when it comes to data and AI use, but also when it comes to tax administration and tax, it's important that the human aspect means to be maintained. And then also what I got from the conversations is implementation of AI is very costly, and every tax administrations are at a different level or stage of implementation. But then, of course, as we are implementing AI, it is important that data and taxpayer confidentiality is is very important. So it is important that the practical guide considers all of those balances, you know, and making sure that when it comes to implementing an effective AI in tax administration, all of those bases are covered. Just specifically to Angela, I just want to say thank you for your contribution. You know, you talk about generative AI and I want to say thank you for Singapore's case study and example, which we will use in the system. Enrique, I just want to say again, thank you for your contribution. You know, we, as you heard from the contributions, you know, AI implementation, especially when it comes to developing countries, we even have some tax administrations in developing countries that haven't even automated their systems. So it has to be gradual, and I'm quite sure we will get to the level that you were just talking about in due time and at different stages based on the country and the level at which they are implementing their tax AI implementation. Shines, thank you so much for, and like everybody else, the human side. You know, we cannot remove the human aspect when it comes to digitalization, when it comes to tax administration, it is important that we continue to keep the human side to the implementation of AI. But I just want to say thank you to each and every one of you for your contribution. Alexander, if you can come in and just maybe add.

Thank you. Thank you, Genevieve. I also know that on the paper that the Human factor is a key element in that, and it could be considered in three dimensions, like human rights, decision-making process still should be on the side of the human, and of course, needs for skilled human resources in the environment when especially developing countries, they do not have enough funds in order to implement AI, taking into account not only the technology, but skilled personnel. And again, I would like to thank everyone who provides his or her comments, and these comments will be taken into account and will be considered during the future work. It's only outline of our deliverable, which we will present next session as a text, as a draft, the first draft of this guide. And I would like again to thank subcommittee members who provided their valuable input. And especially, I would like to thank subcommittee members, some of them are here, and they also I believe, take some notes from the discussion we have, we have here, so thank you very much, and we will try to do our deliverable as practical as possible, thank you.

Thank you very much, Alexander. Just quickly, I want to just join Alexander to say a big thank you to our subcommittee members who are here present. We have a large number that actually joined us in person, and I think we have a few that are also following us online. So I just want to say a big thank you for joining us.

Thank you very much, Jeneba, Alexander, and Lorenzo for presenting the paper, and we look forward to receiving the first draft chapters in the next session. Thank you very much. And by that, we will break for lunch. It's one o'clock, and we will resume with the paper on indirect taxation at three o'clock. Enjoy lunch, everybody.