Artificial Intelligence and emerging technologies in the context of international peace and security, with emphasis on maritime security aspects

His Excellency the Minister of Foreign Affairs of Greece, Mr. George Gerapetritis, and His Excellency the External Affairs Minister of India, Dr. S. Jaishankar. We are privileged to have with us two distinguished UN briefers, Under-Secretary-General and Special Envoy for Digital and Emerging Technologies, Amandeep Singh Gill, and Under-Secretary-General and High Representative for Disarmament Affairs, Izumi Nakamitsu. I would also like to recognize our cooperation with the Security Council report. in preparing today's event. I warmly welcome its Executive Director, Ms. Marla Candia Thompson, and Mr. Rodrigo Saad, Policy Analyst, whose presentation of Security Council Report's policy paper will provide an important contribution to our discussion. Ladies and gentlemen, we are all witnesses to how artificial intelligence is transforming the tempo of our world. Its extraordinary potential is already apparent, but so too are the profound questions and risks that accompany a technology developing at unprecedented speed in a world where conflict now ignite in any instant and crisis gather force before diplomacy has found its voice. At sea. emerging technologies and artificial intelligence can lift the veil over the vast expanse of the seas and the oceans. They offer new possibilities to enhance maritime domain awareness, early warning and the protection of critical maritime infrastructure. Yet, the same technologies can also introduce new vulnerabilities and risks, reminding us that innovation must always be anchored in international international law with UNCLAWs at its cornerstone. The importance and full applicability of international humanitarian law and international human rights law must also never be overlooked. Today's discussion will allow us to explore both the opportunities these technologies offer and the challenges they present from different perspectives. Without further ado, it is my pleasure to give the floor to His Excellency, Minister Yerasimopoulou. Mr. Minister, you have the floor.

Thank you, Ambassador. Dear colleagues, dear friends, it's a great honor to welcome you today to the joint side event of Greece and India on artificial intelligence, emerging technologies, and international peace and security. I am particularly pleased to co-chair this discussion with my good friend, the Minister of External Affairs of India, Dr. Jaishankar, who's one of the most distinctive personalities in international diplomacy. Let me also warmly thank our distinguished briefers, Amanatip Singh Gill, Izumi Nakamitsu, and Rodrigo Saad. And I would like also to acknowledge the presence of Security Council's Reports Executive Director, Samala Kandian-Thompson. We are going to also discuss the Security Council report, Artificial Intelligence, Geopolitics and the UN Security Council, especially examining the implications of artificial intelligence on peace and security, especially in the maritime domain. Ladies and gentlemen, artificial intelligence is already transforming the international security environment. It offers significant opportunities to strengthen our ability to prevent conflict, protect civilians, and support peace. At the same time, it creates new vulnerabilities and raises profound questions about responsibility, accountability, and the application of international law. The question before us is therefore no longer whether AI will affect international peace and security. It already does. The question is how we choose to shape that transformation. For Greece, the starting point is clear. Artificial intelligence must remain human-centered, responsible, and firmly anchored in international law. The UN Charter, international humanitarian law, and international human rights law remain fully applicable as technologies evolve. And one principle should remain constant. Technology can augment human judgments, but it cannot replace human responsibility. Important international initiatives have already contributed to this discussion, including the RAIAM process, the political declaration on responsible military use of artificial intelligence and autonomy, and the India AI Impact Summit earlier this year. Greece has also sought to advance this conversation with the United Nations. In April 2025, together with France and the Republic of Korea, we convened an ARIM formula meeting of Security Council on artificial intelligence and international peace and security. Today, we want to take this discussion further, with particular attention to the maritime domain. The opportunities are considerable. AI can strengthen maritime domain awareness and early warning. contribute to safer navigation and shipping and enhance the protection and resilience of critical maritime infrastructure. More fundamentally, it can help us move from reaction towards anticipation, identifying emerging threats earlier, understanding them better, and responding before they escalate. But AI is also inherently dual use. Capabilities developed for legitimate and peaceful purposes have indeed vulnerabilities affecting port, shipping, critical infrastructures and undersea cables. The maritime domain therefore provides a compelling illustration of both the promise and the risk associated with artificial intelligence. I would also like to acknowledge India's long-standing contribution to discussions on science and technology in the context of international security through its annual initiative in the first committee this morning. We had first gathering of the Group of Friends of Safety and Security of Shipping and Seafarers, where my good friend, Dr. Jay Sanka, led the initiative alongside with the Foreign Minister of Liberia. For Greece, maritime security has been a central priority throughout our tenure on the Security Council. We have sought to keep the protection of maritime routes and critical infrastructure, and above all, the safety and security of seafarers, firmly on the Council's agenda. As Greece prepares to assume the presidency of the Security Council in October, we believe that the Council should look not only at the risk created by new technologies, but also at how they can strengthen our collective capacity for prevention and peace. Let me highlight in this respect four considerations. First, we should harness AI as a tool for prevention. AI can strengthen early warning and conflict analysis, support peace operations, facilitate mediation and humanitarian response, and improve maritime situational awareness. The ambition should not simply be to manage crises more efficiently. It should be to anticipate them earlier and, where possible, prevent them. Secondly, we need stronger partnerships. Governments and international organizations do not possess all the technology or expertise required to understand this transformation. We need stronger cooperation with technology companies, the scientific community and academia based on clear responsibilities and appropriate safeguards. When it comes to maritime security, we need to also engage with the maritime community. Thirdly, our multilateral institutions must keep pace with technological change. The Security Council should be able to consider the implications of emerging technologies more systematically when they affect international peace and security, drawing more regularly on scientific and technical expertise. And finally, innovation must go hand in hand with responsibility. And this is the most critical factor of my presentation. The opportunities offered by AI cannot come at the expense of international law, human oversight and accountability. There are no obstacles to innovation. They are conditions for trust. And the truth is that the artificial intelligence is not another phase of industrial revolution. It's a new industrial revolution. And the idea is that for the first time, I think there is a direct challenge towards the supremacy of human brain. The positive things we all understand and know, health, education, growth applications, obviously addressing global challenges such as climate crisis. However, there is a significant risk and this comes through the possible under regulation or unchecked implementation of artificial intelligence. First of all, the impact on vulnerable groups and especially children cannot go unattended. Secondly, massive job destructions. We expect to have a transformation of the labor market as AI proceeds immensely. Thirdly, concentrated wealth. Obviously, since AI will develop technologies that will be at large scale marketable, it's obvious that this will produce profits. And in this respect, we can expect global inequalities already existing at large scale to become greater. Now, we shouldn't, in my eyes, have an illusion that artificial intelligence operators can indeed self-regulate. This is something that I cannot believe and expect. We need strong regulation when it comes to AI. Strong regulation concerning license, concerning supervision, concerning a sanction system for possible violations of fundamentals concerning the implementation of artificial intelligence, and most importantly, a very powerful international regulator to actually see how artificial intelligence progress to the benefit of humanity and not to the benefit of specific humans. Technology will continue to evolve. The question for us is whether our institutions, our cooperation, and our collective ambition can evolve with it. And this is a conversation we want to advance today, and one that Greece and India intends to carry forward within the United Nations. Thank you.

Thank you, Minister. I kindly invite Dr. S. Jaishankar, External Affairs Minister of India, for his opening remarks.

Thank you. Co-Chair and my good friend, Minister Gerapetritis, Under-Secretary-General Amandeep Singh Gill, Under-Secretary-General Izumi Nakamitsu, Mr. Rodrigo Saad, Excellencies, distinguished colleagues. I'm honored today to co-chair this discussion with Foreign Minister George Gerapetritis on a subject that is both increasingly consequential and rapidly evolving. Artificial intelligence is transforming our economies, our societies, and inevitably, our security environment. The question before us is therefore no longer whether AI will have implications for international peace and security, but how we harness its possibilities, anticipate its consequences, and mitigate its risks. This requires balance. An approach driven exclusively by apprehension risks constraining innovation and denying its benefits, particularly to developing countries. But An approach that ignores the security implications of AI would be equally short-sighted. This February, we hosted the AI Impact Summit. The New Delhi Declaration was premised on a simple proposition. The promise of AI can be realized fully only when its benefits are shared by humanity. Its emphasis on democratizing access to AI resources building secure and trustworthy systems, investing in human capabilities, and using AI for economic growth and social good reflected the concerns and aspirations of countries at very different stages of technology development. At the same time, we must also recognize that AI has a clear security dimension. It can improve early warning and situational awareness, It can strengthen peacekeeping and humanitarian operations. It can support conflict analysis. It can enable faster and better informed decisions. But it can also amplify cyber threats, facilitate disinformation, enhance autonomous capabilities, and lower barriers for malicious actors. Now, nowhere is this duality more evident than in the maritime domain. Recent conflicts have demonstrated how disruption at critical maritime choke points have consequences that are far beyond their immediate region. Attacks on commercial shipping affect not merely the parties to the conflict, but food security, energy security, and economic stability across the world. Artificial intelligence can enhance maritime domain awareness by processing large volumes of satellite, radar, and vessel tracking data. It can improve anomaly detection, assist search and rescue operations, strengthen disaster response, identify suspicious maritime activity, and enhance the resilience of ports and other critical infrastructure. Our Information Fusion Center for the Indian Ocean region, which is located just outside Delhi, works with partners to strengthen maritime domain awareness and information sharing. But as I pointed out, these same technologies also create vulnerabilities. AI can be used for sophisticated reconnaissance, vulnerability mapping, and cyber operations against ports, vessels, communication networks, and undersea infrastructure. Increasing autonomy in maritime systems raises questions of attribution, accountability, and escalation, while also compressing decision-making time. We therefore need an approach that preserves human responsibility and accountability. International law, including the UN Charter, and the laws of seas as reflected in UNCLOS remain the essential framework governing state conduct. Their application must be consistent with existing international obligations. These considerations also underlined by AI governance cannot be reduced to regulation alone. We need greater transparency, confidence building, capacity development, technical cooperation, and exchange of best practices. Developing countries must have the capability to assess AI systems, protect their infrastructure, and participate meaningfully in the shaping of norms. Friends, colleagues, the United Nations has an important role in building that common ground. Different UN bodies have distinct mandates and expertise. Discussions on AI should reinforce rather than duplicate them. In the Security Council, the focus should appropriately remain on those applications and consequences of AI that have a demonstrable bearing on international peace and security. India will bring precisely this approach to the Security Council if elected for the 2028-29 term. As I stated when we launched our candidature earlier this year, We seek a world where the promise of technology is fully realized while safeguarding against its misuse and application, and where our oceans remain safe and secure so that maritime lifelines are not threatened. AI and maritime security will therefore both be important priorities for India during our Security Council term. We would seek to promote practical discussions focused on building capacities sharing experiences, strengthening resilience, and developing a clearer understanding of emerging risks, rather than creating new divisions or imposing prescriptions. Our objective should therefore be to ensure that AI strengthens human security rather than undermines it, that it expands opportunity rather than inequality, and that it helps to make our shared maritime commons safer rather than more contested. So I look forward to hearing the perspectives of our distinguished briefer and to a very productive exchange this afternoon. Thank you very much.

Thank you, Mr. I now invite the Under Secretary General and Special Envoy for Digital and Emerging Technologies, Mr. Amarjeet Singh Gill, for his briefing. You have the floor.

Thank you very much. Good afternoon, everyone, Your Excellency. Minister of Foreign Affairs of Greece, Your Excellency Dr. Jaishankar, Minister of Foreign Affairs, Minister of India. It's an honor to be convoked to this meeting, co-hosted by two nations with a great maritime tradition. I will confine my remarks to the non-military aspects of this issue, given, Minister, your exhortation that We have these different mandates, different responsibilities, and we need to work together in that spirit. And I'm sure Izumi will get into that aspect. Now, recently in July, I had the opportunity of being on a fully autonomous yacht. And this was right after a ride in a fully autonomous flying drone. So these technologies are no longer science fiction. They are a part of daily life. And there's tremendous potential in the maritime domain, parking ships safely and in a fast manner in congested ports, the safety and security of seafarers, sustainability of fossil fuel use, and many, many other areas in the maritime space will benefit, I'm sure, from AI. Likewise, there are risks, even though in the maritime domain, the risk of collateral damage is sometimes more manageable than it is on land, but still there are plenty of risks that we have to contend with. There are in the non-military domain, risks around counterterrorism, cybersecurity, other areas where non-state actors are particularly active. And we've seen that in many maritime areas. Experience with piracy on the high seas is just one example. So if you look back at the Global Digital Compact, the the focus is on the non-military domain, but there is this overlap with peace and security issues. And the Secretary-General's attempt has been to, through his own office and otherwise, to bring a whole of UN perspective to these issues, because technology doesn't sit in one place, it travels, it affects different domains. And this technology in particular is not just dual use, but it is repurposeable in different ways. And it is combinatorial, can come together with other technologies for novel effects. So what is happening at the UN is this coordinated approach, part of which is improving our own work inside the UN, UN 2.0, using AI and other emerging technologies for greater efficiency, better delivery of humanitarian action on the ground, better delivery on our mandates given by member states. And then there is an external facing role which has a policy dimension. And you see that illustrated in the work of the new mechanisms that have been created through the Global Digital Compact, the UN's independent international scientific panel on AI, 40 global experts from around the world who are giving regular independent assessments of the direction, the pace, the opportunities and the risks related to this technology. On 1st July, in their preliminary report, they said that the possibility of these autonomous agents acting beyond human, explicit human instructions could not be ruled out. And just 10 days later, that was borne out by this incident where these agents escaped their sandbox and compromised the cybersecurity of a significant platform. And now we see there have been other such incidents. So the panel's work is important. And earlier this week, early morning on Monday, 5 a.m., they released a thematic brief focused on this particular incident and what it implies for multilateral cooperation. And this resonated at the Security Council as well. just a couple of days ago, when many of the tech companies involved in these developments, including the victim and the perpetrator of this attack, brief the Security Council and in fact, in a somewhat surprising turn, asked the international community to come together and address these risks as a matter of priority. The Security Council was also briefed by the co-chair of the independent panel, Dr. Joshua Now, going forward, there is another meeting of the panel that will happen at the end of November in Barbados, and there is a request from the panel to convene with leaders from these frontier model companies, the representatives of AI safety institutes and other independent experts. Minister, you mentioned the importance of capacity around safety testing should be available everywhere. So this discussion would move along. given the urgency of the issue. And next year, there'll be the global dialogue, second global dialogue on AI governance in May in New York. I think there are three areas, and I want to close with this, three areas where we need to connect the dots and make sure that multilateral cooperation is focused on those three areas. One is the travel from the evidence to the policy. So the panel's remit is policy relevant, but not policy prescriptive advice, and Member States have to take forward the policy actions in the dialogue and in other forums in the multilateral space. The other bridge we need to cross is on capacity. So science and policy are important, but without capacity, they remain abstract. So capacity is not only a development imperative, it's also a governance imperative. So you need these assessment capabilities, even whether you buy, borrow, or build AI, you need those capacities. So we need to translate from the AI divide to this capacity lift. And finally, I think we need to make the connection to the existing bedrock of norms. International humanitarian law has been mentioned, IMO conventions have been mentioned. So there's no need to reinvent the wheel. Some of the suggestions made in the Security Council, for example, if you rely on the bedrock of existing norms, then you can make the translation to this domain seamlessly without duplication and without taking away sovereignty from national implementation authorities, national regulatory authorities. So we look forward to working on those areas of urgent importance. Thank you so much.

For his most informative briefing and I invite the Under-Secretary-General and High Representative for Disarmament Affairs, Ms. Yumi Nakamitsu, for her briefing.

Thank you very much, Madam Ambassador. Excellencies, distinguished guests, ladies and gentlemen, dear colleagues, it is indeed a privilege to join you today to discuss artificial intelligence in the context of international security, with an emphasis on maritime security. And I would like to extend my sincere gratitude to their excellencies, Minister Gera Petricis and Minister Jayashenka, for their leadership in convening this very timely side event. It is, of course, no longer news to anyone that AI affects everyone and every aspect of our lives, and it will continue to be the case. International peace and security, of course, is no exception. Militaries and increasing militaries are increasingly integrating AI across a wide range of functions. the accelerated pace of technological development must be matched by effective governance, safeguards, and international cooperation to ensure that human agency, judgment, and control are preserved as these systems become more agentic. Specifically in the context of maritime security, AI presents both opportunities and, of course, risks. So let me begin with the opportunities. AI can improve maritime domain awareness by analyzing large volumes of information from satellites, radar, sonar, and other maritime tracking systems. For states responsible for large maritime areas, These capabilities can support more effective monitoring and improve the detection of illicit or suspicious activities. AI can also support search and rescue, early warning, monitoring and verification, and the implementation of relevant Security Council measures as well. Other emerging technologies may strengthen these applications. Advances in position, navigation and timing, for example, could help AI enabled and autonomous systems operate when satellite based signals are disrupted and unavailable. But the same technologies could also introduce new security risks. Let me mention a couple of them. First, Uncrewed surface and undersea vessels are proliferating rapidly among militaries, changing the character of maritime operations. AI is particularly attractive enabling technology for these vehicles, allowing them to operate with greater autonomy, over longer distances and for longer periods, including in environments where communications may be limited or disrupted. Greater autonomy and speed may reduce the time available for human assessment and intervention, while interactions between AI-enabled autonomous systems could increase risks of misperception, miscalculation, or unintended escalation. This makes it particularly important to preserve human judgment and control over the use of force. Second, AI is changing what can be observed at sea. As the ability to detect, classify, and track military activities improves, states may reassess the vulnerabilities of their own forces. This becomes particularly consequential where nuclear weapons are involved. These developments, whether actual or perceived, have the potential to undermine strategic stability and create use it or lose it mentalities about nuclear weapons, raising the prospect for first strikes. More broadly, the potential integration of AI into early warning systems or nuclear weapons systems create additional risks. The former increases prospects for spoofing, hacking, or misidentification, leading to increased prospects for mistake or miscalculation and escalation. The latter, that is the nuclear weapons systems, The Secretary General has been very clear. Humanity's fate must not be left to an algorithm. Pending their total annihilation, humans must remain in control of nuclear weapons at all times. Now, third, AI-enabled maritime systems depend on an increasingly complex digital ecosystem. This creates dependencies and potential vulnerabilities. Submarine cables are a load-bearing part of this digital infrastructure. Recent incidents in which such cables were cut, whether accidentally or on purpose, have highlighted both their importance and their vulnerabilities. In addition, global navigation satellite system, GNSS, Jamming or spoofing incidents have already been documented in the Black Sea, Persian Gulf, and around contested waters in the Asia-Pacific region. These opportunities and risks point to the same conclusion. AI becomes more deeply integrated into maritime domain, governance and safeguards must evolve alongside technological capabilities. Though its two resolutions on the through its two resolutions on AI in the military domain, the General Assembly has actually begun a process that has thus far yielded a report of the sector general in 2025 last year and a fruitful informal exchanges earlier this year. At the same time, the efforts of Greece and others have shown there is a role for the Security Council as well. At a minimum, this entails keeping abreast of technological developments and their impact on Security Council mandated sanctions regimes and on situations on its agenda. Given the Security Council's primary responsibility for international peace and security, it could also play a more strategic role should it so choose. There has been a great interest among Council members to regularize engagement on emerging technologies, including artificial intelligence. The UN Office for Disarmament Affairs is always available to support such efforts. We are navigating turbulent times. Military expenditure is increasing. Mistrust is spreading. and the risks of miscalculation and escalations are growing. Rapid technological challenge adds another layer of complexity to this environment. So together, we can ensure that advancements in science and technology serve rather than undermine the purposes and principles of this organization and, of course, the humankind overall. I trust that today's discussion will be a positive step forward. I'm looking forward to the exchanges that will follow. Thank you very much.

Thank you, Under-Secretary-General Nakamitu, for your very pertinent briefing. And now, I would like to introduce the Executive Director of the Security Council Report, Ms. Amala Candia-Thompson, for her briefing. Thank you.

I get to do a short introduction, whereas my colleague will have to do the bigger briefing. But, Excellencies, distinguished guests, dear colleagues, on behalf of Security Council Report, I would like to thank the governments of Greece and India for organizing this event today. It is an honor for us to be able to launch our policy paper at this important event, and I would particularly like to express our gratitude to Greece for supporting the paper. For those of you who may not be familiar with SCR, we are an independent and impartial think tank dedicated to the transparency and effectiveness of the Security Council. For over 20 years, we've provided information and independent analysis of the Council's work through our regular publications, as well as in-depth reports. Our research reports have tended to focus on thematic issues that are significant in the work of the Council. And as the international peace and security landscape evolves, we have begun devoting greater attention to emerging challenges that we see affecting the Council's work. In March, we published a research report that examined the Council's engagement with new and emerging technologies, including information and communication technologies and AI, and identified options for strengthening engagement in these areas. Our policy paper, which is titled, "Artificial Intelligence, Geopolitics, and the UN Security Council," builds on that research, providing a more focused examination of AI's implications for international peace and security, and the council's role in addressing it. I'll leave it to my colleague, Rodrigo, who was the author of this paper and our research report, to provide you with the details of the paper. but I hope that the policy paper will provide a useful resource for Council members and the wider UN membership for continued engagement on this rapidly evolving issue. Thank you.

And now I give the floor to the author of the report, Mr. Rodrigo Saad.

Thank you, Madam Ambassador, Excellencies, distinguished delegates, dear colleagues. I would like to thank Greece and India for convening this event and Shyamala for her kind introduction. We tend to think of artificial intelligence as something that lives in the cloud, but in practice, it rests on the sea. The minerals, the semiconductors, and the computing equipment that AI depends on travel by ship. The data that connects AI systems across borders travels through cables on the ocean floor. The relationship runs in both directions. AI is changing what states and non-state actors are able to do at sea for those who protect the maritime domain and for those who threaten it. That is why the paper we are launching today, Artificial Intelligence, Geopolitics, and the UN Security Council, runs maritime security through it as a thread to show how AI intersects with issues already before the council. Greece and India have each brought maritime security before the Council at the highest political level. Greece last year in a debate chaired by its Prime Minister and India in 2021, when the Council adopted its first product devoted to maritime security. In 2024, the Council committed to taking scientific advances into account more systematically, where appropriate and within its mandate. So my question today is a practical one: how can the Council build on those commitments? To answer that, we have to start with the competition now surrounding this technology. Expectations about where the technology is heading are shaping investment and strategic decisions today. That includes the prospect of systems matching or exceeding human performance across a broad range of tasks, what is often called artificial general intelligence. And this competition extends well beyond the technology itself. Advanced AI depends on semiconductors, computing infrastructure, energy, data, and critical minerals. And these resources and capabilities are unevenly distributed. Many developing countries supply essential minerals, while advanced computing infrastructure and model development remain concentrated elsewhere. Resource producing countries are seeking a greater share of the value created, the chance to develop their own capabilities, and a voice in how AI is governed. This has consequences for peace and security. In fragile and conflict-affected settings, rising demand for minerals can reinforce conflict economies where extraction finances armed groups. Disputes over access, ownership, and revenue sharing can shape the incentives of conflict parties and prolong hostilities. It also runs through the maritime domain. More than 80% of international trade in goods by volume travel by ship. And where production is already concentrated, a disruption at a single choke point compounds the exposure. Beneath the sea, some 500 cables carry more than 99% of international data traffic. Three or four are damaged somewhere in the world every week, mostly by anchors and fishing gear. But recent incidents have raised concerns about deliberate damage. But AI is already shaping peace and security in more immediate ways, and not only for the worse. The UN is applying AI to its peace and security work. It has used AI-enabled digital technologies to consult large numbers of people in their own languages, including in support of peace processes. Missions are also using AI-enabled tools to analyze online discourse and radio broadcasts for harmful narratives directed at civilians and at the missions themselves. Situational awareness platforms such as UNITE Aware bring mission data together into a single operational picture. UNITE Aware was launched in this building during India's Council presidency in August 2021. At that same meeting, the Council recognized that technology can act as a force multiplier in peacekeeping for the safety of peacekeepers and for the protection of civilians through better early warning and more timely decisions. AI could extend that commitment. Applied to the data missions already hold, it could help assess where violence is likely to occur and inform patrol planning. Similar capabilities could support mediation, ceasefire monitoring, sanctions monitoring, and conflict analysis, much of the Council's existing toolkit. The effective use of these tools depend on the quality of the data missions record. They depend on the personnel able to judge the outputs, including when the outputs are wrong. And they depend on protecting sensitive information. A decade ago, a UN expert panel found that UN peacekeeping was, quote unquote, "well behind the curve on technology" and introduced the idea of a digital peacekeeper. The question now is whether we are moving towards something closer to an agentic peacekeeper, systems that can automatically monitor sources, cross-check claims, and keep assessments current. But this is not a substitute for personnel. I would make the opposite point, in fact, to take the repetitive work off them so that judgment, political analysis, and relationships with local communities receive more of their time. and humans must retain meaningful oversight of these tools. In corporate evaluations, misaligned model behavior has done limited damage. Inside a peacekeeping mission, the damage would not be so contained. The capabilities that support this work are also becoming available to actors seeking to undermine it. Generative AI lowers the barriers to activities that once required expertise, greater expertise, and resources. UN sanctions monitoring has documented terrorist groups using AI to produce and translate propaganda and to seek guidance on improvised explosive devices and target selection. Just this month, Anthropic reported that threat actors in northern Yemen had used its models across three guided weapons programs. The actors test fired a guided rocket, and when it failed, they went back to the model to diagnose the fault. The same report describes blocked attempts to use the models for research that could support biological weapons development, a concern squarely within the remit of the Council's 1540 Committee. The same dynamic runs through cybersecurity. Capable coding models and AI agents now let a single operator run campaigns that until recently required a full team. In that same anthropic report, one state-linked group ran an AI-driven espionage operation against a government agency responsible for maritime shipping and tracking. The same capabilities work for defenders. AI can find and fix vulnerabilities as quickly as it can exploit them, which matters most for countries with the least cybersecurity capacity. Generative AI is also changing the information environment in which diplomacy and conflict unfold. Fabricated images, video, or audio could suggest that a ceasefire has been broken or that a leader has ordered an attack. Such material circulates before authenticity can be established, potentially fueling violence and undermining mediation. Then there's military decision making. AI is being used to sift intelligence at volume and to help identify targets. That compresses the time commanders have to verify, and under pressure, people tend to trust what the machine tells them. Just last week, CNN reported that earlier this year, an AI-assisted intelligence assessment concluded that a foreign vessel in the Middle East was carrying components of a nuclear weapons program. And according to four sources, the military prepared to board the ship. Aircraft were already in the air. Only then did officials examine how the assessment had been produced and found that a chatbot had misidentified the cargo. One source said the report almost started a war. That operation was a maritime interdiction. The delegation of decisions involving lethal force raises further questions. Autonomous weapons are already in use, where systems select and engage targets without further human intervention. Questions of accountability and compliance with international humanitarian law follow. Peace operations must work in environments increasingly shaped by AI amplified risks, cyber attacks on missions and host country systems, drones and autonomous weapons that endanger personnel, disinformation that erodes trust with local communities and complicates mandate delivery. Missions need the capabilities to operate safely in that environment. A further concern raises from the behavior of these systems themselves. During evaluations, AI agents have bypassed restrictions, attacked real third party systems, and concealed what they were doing. Two days ago, Australian authorities disclosed that in June, an OpenAI agent had hacked the government health statistics portal, accessing non-public files. This is the first known case of an AI agent breaching a government system. The company said that its model had taken unintended actions and notified the Australian government three months later. More capable agents behaving this way could cause serious harm. If one gained access to critical infrastructure or to control over physical systems of vessel at sea or port, the consequences would not be confined to the digital environment. These AI systems are also increasingly used to build the next generation of models, a process recursive self-improvement. If that continues, capabilities may advance faster than we can understand them, and meaningful human control becomes harder to maintain. That is why testing, independent oversight, and incident reporting matter now, and why on Wednesday the companies building these systems came to the council asking for rules. But competition between companies and between states rewards speed over caution. Arrangements of that kind need the wider international community. The Council's contribution should be grounded in its mandate. There are political differences about how that contribution should go. Many favor more sustained engagement. Other questions whether broad discussions of AI belong in a body of 15 members and caution against treating AI primarily as a security issue. That disagreement was audible in the Council chamber as recently as Wednesday. Those differences make agreement on new formal mechanisms unlikely in the new term. They also make complementarity with the wider UN system more important. But the technology will continue to advance, and so will its effect on peace and security. The Council's engagement thus has to advance along with it. Two things make that difficult. The first is sustained attention. There's no dedicated agenda item on AI and no standing reporting cycle. The Council has held four formal meetings on AI to date. Each of those happened at the initiative of individual members. But the presidency changes every month, and elected members leave after two years. Nothing reliably carries the analysis across those transitions. The second gap is access to information. The expertise exists within the UN system in missions, sanctions panels, counterterrorism bodies. as well as among researchers and practitioners outside the UN. What members lack is a standing form that brings it together at the point where it bears on decisions that the Council is taking. The main option the paper puts forward to address these gaps is an informal expert group on technology, peace and security. Let me describe exactly what that would entail. Council experts meet out of public view a few weeks before a vote on a draft resolution or a mandate renewal. They are briefed by people who know the subject from within and beyond the UN system. They get to ask the briefers questions. They work out what the technology means for the situation in front of them. And then they take that back to their delegations before negotiations begin. The Council has done this before with issues that once sat at the edges of its work. The Expert Group on the Protection of Civilians was created in 2009 at the initiative of the UK and meets ahead of mandate renewals with the protection dimension. The Group on Women, Peace and Security followed in 2016. In both cases, the purpose was the same, to take a cross-cutting thematic issue that was raised occasionally and make it a routine part of the Council's work. That is a task with technology now. Take maritime security. Such a group could bring sanctions, sanctions experts together with specialists in maritime monitoring and ask how AI might help identify illicit activity at sea and how it might help conceal it. Those answers bear directly on mandates on on the council's the council's renews. And it would give the council a link to the wider system. Members could invite the co-chairs of the independent international scientific panel on AI to brief them. On Wednesday, they heard from one of the co-chairs, an expert group could make that regular rather than exceptional and help the council complement the work rather than duplicate it. And because it would meet regularly, analysis would survive the change of presidency and the change of membership. The same forum could also host forward-looking discussions before technological developments produce a crisis. Interested council members could set this up themselves without a resolution or council product. Two further steps could reinforce this work. Interested members could adopt voluntary shared commitments, coordinating across successive presidencies so that AI-related issues raised under one are carried into the next. And the Secretariat could prepare an aide memoire on technology, bringing together the Council's existing agreed language and practice. That is also not a novel idea. The protection of civilians has had one for years, and it is what gives the expert group something concrete to work from. Excellencies, consider the headlines from the past week alone. A rogue AI agent hacked a government database. An incorrect AI assessment nearly put armed personnel aboard a foreign vessel. And the heads of AI companies briefed the Council, warning about the risks of losing human control over increasingly capable systems. This was one week. The Council has met formally on AI four times in over three years. More systematic consideration of the rapidly evolving threat landscape is warranted. The Council has committed to taking scientific advances into account more systematically. That is within reach, and it does not require anyone to agree on the broader questions of AI, only to accept that it is already shaping the situations before the Security Council. Thank you. Thank

you, Rodrigo. and introducing this very important report that was authored by you. And now I will open the floor for discussion and I will give the floor to the Minister of Foreign Affairs of Romania, Ms. Oana Toiu. You have the floor.

Thank you. I first of all want to say on behalf of Romania that we are grateful for the leadership of Minister Jaishankar and Minister Gerapetritis. We've spoken at length throughout this week about the risks that we see on maritime security. And I think specifically this conversation is what we need to be able to move forward on the instruments that in the present moment and in the future are needed so we're able to secure that. In the deteriorating geopolitical environment in the maritime sector, it has become increasingly shaped by the dynamics of hybrid warfare. We see an increase in combining military activity with non-military means, especially under the context of the Russia's war of aggression against Ukraine. We have noticed that our borders in the Black Sea, specifically, a lot of hybrid interferences. And that has led to an unstable environment in maritime security, one that doesn't just affect our countries and the countries of the Black Sea, but we have seen them disrupt the food and energy security, especially as the Black Sea is a key corridor for shipping grains. Romania is the number one grain exporter of the European Union. But of course, here on the global scene, it's very important to discuss as well. about Ukraine's grain capacity to reach the countries that are in need. And for that, maritime security is essential in the Black Sea. Now, concerns about the impact of AI in the military domain, they gain ground in various fora and institutions. And it's important that we do have the officials, industry, and experts from General Assembly and UN buy this to rein the interest growth about fostering norms and opera. lies responsible AI critical domains stretching from understanding to risk of application. The rapid evolution of technological development influences all fields of strategic importance, including the maritime sector. Now we have emerging information technology and AI. They are potentially beneficial for the rapid response that we see. In our case, we are together with Bulgaria, building the concept for a EU maritime security hub to be able to have an enhanced awareness and the situational awareness of what's happening. We want to increase the partnership we have with other countries across the world on maritime security, because we think we need to move fast in sharing practices, in sharing know-how to be able to face the challenges that we see. We also are keen on discussing more on protecting the critical infrastructure. in the Black Sea, not just the current infrastructure, but the potential to develop more of these connections we see as being very important. And yes, here we're speaking about maritime surveillance, we're speaking about cybersecurity, and here on cybersecurity, Romania is also host to the European Competence Center on Cybersecurity, and we have advance. We've launched last year at the Munich Conference, the cybersecurity partnership between Romania, Ukraine, and Moldova in the Odessa Triangle. I'm mentioning these three countries because we have faced similar risks and an intense volume of challenges on cybersecurity. So we put together the lessons that we have and the sharing between our experts to see how we can answer to those. Of course, pipelines, some marine communication cables, offshore energy projects. In our case, again, we're the number one natural gas producer in the European Union, and we have an 80% increase plan for next year with the Neptune Deep natural gas project in the Black Sea. But we are seeing increased activity around that perimeter, and we do need better ways to monitor that and to be able to deter any risks on it. In the context of strengthening the security of the European Union and implicit of its member states, the use of advanced AI represents a strategic pillar for the optimization of naval space surveillance, efficient border management, and prompt counteraction of emerging asymmetric threats. Now, I'm mentioning the European Union to speak a bit on behalf of my colleagues in the European Union that are not here today, because we have launched a new maritime security strategy and The AI and technical part is very important for it. Frontex, the European approach, makes the use of AI expert tools and algorithms to neutralize the risks associated to ships that have hidden identities. As you know, for us in the sanction package, the managing of the shadow fleet is very important and access to data to be able to do that. To provide early warnings. regarding unmanned surface vehicles. Now, we have seen maritime drones with explosives in extreme closeness to our ports. Our teams have neutralized them, but this is extremely risky because there were cases, for example, when those drones were Ukrainian, obviously with no intent on the Ukrainian side. just being autonomous vehicles, they have lost connection to them at some point on the Black Sea. And of course, that creates a different new level of risks, even under false flag type of activities or simply we need better instruments to be able to manage the perimeter. Now, this being said, the ever growing use of AI in the decision making process requires a strict management of cybernetic vulnerabilities and operational dependencies and the balance between AI use and resilience. It is very clear, I think, that we need these instruments going forward to be able to have better analytical data and to have a better response. But I also think it is essential, and that's why we welcome this conversation, that we stay closer together, like-minded countries, in the use of these instruments to be able to adapt faster to what is needed. Thank you.

And now I invite the the Assistant Foreign Minister for Advanced Science and Technology Affairs of the United Arab Emirates, Mr. Omran Sharaf Al Hashemi. You have the floor, Minister.

Thank you very much, Excellencies. I would like to take this opportunity to thank Greece, India, the UN, and the UN Security Council for convening this very important discussion. I'll try to be very brief and not take a lot of your time. It's a very important topic, I believe, that we're discussing today, given the evolving nature of this technology and the rapid development that's happening, combined with the geopolitical landscape globally, I think. It highlights the importance of discussing different topics related to AI when it comes to the security discussions. That includes also maritime. I think when we look at AI in the field of maritime, I think It definitely plays a big role in detecting threats. It plays a big role in mitigating threats and also in eliminating threats. However, at the same time, as was highlighted and we've seen that it also plays a big role in planning threats, executing threats, and causing real damage and harm. So I think it's very important for us also when we look at it, I think we really need to look at it from an angle of dual use. AI at the end of the day. The technology is dual use, but also so is the data used is dual use. So when it comes to governing this very important technology, I think it's very difficult for us to have a blank approach by just having set of standards that applies on technology and in data the same way. And I think this is very important for us and approach that we should take is totally understanding the differences and going into the detail of knowing how to govern these different elements and aspects, which is very, very complicated. But at the same time, we cannot wait for those systems to be in place because there's a lot of benefits also to this technology, given the challenges that we're facing today. So just to keep it very short and going straight to the point, I think it's very important for us to understand that also maritime is about the safety and the security of the sea ferries. It's about ensuring the freedom of navigation and also upholding the international law. But also it's also about connectivity when it comes to the undersea cables. That is an important angle to take into account. So we have the kind of the traditional economic ecosystem that maritime plays a big role in it, but also now the new and the digital ecosystem, the economic ecosystem. So this topic, I think, is very complicated and approaching it has to be done in a very, very careful way. We should not choke and overregulate, but also we should not underestimate the threats. Thank you very much.

For continuing our discussion, I would like to thank for their presence Minister Jaishankar of India and Minister Arabetidis. And I allow them to leave to continue their heavy schedule. Thank you, Ministers. And now I would like to give the floor to the Director General of the Ministry of Foreign Affairs of Latvia, Ms. Inga Skujina. Madam, you have the floor. Thank you.

Distinguished ministers, colleagues, we thank Greece and India for organizing this important and timely discussion. Today, we are witnessing attempts by some actors to sidestep or even breach international law. No doubt such steps are endangering maritime trade, communications and environment. It has been already mentioned. Among these threats, we must include also the shadow fleet operating outside any safety parameters to unlawfully evade sanctions. Another key risk stems from sabotage against the subsea cable infrastructure worldwide, as it was also just mentioned by the Minister. Emerging technologies add an additional layer of complexity to these risks. Therefore, we welcome the research presented today, which looks at these impacts in a systematic way. As highlighted in the report, artificial intelligence enabled platforms can enhance safe navigation by providing each early warning of potential risks. Yet, artificial intelligence can also be used to catalyze cyber attacks capable of paralyzing ports and navigation systems. Also, the growing use of autonomous and remotely operated ships offer further benefits, while also introducing new challenges. Overall, the intersection between artificial intelligence and the maritime domain brings together two vast spheres. that are not confined by national boundaries. Used responsibly, they offer the potential to connect the world. Used with malign intent, they can produce the opposite effect, disrupting freedom of navigation, communications, and trade. To put it in a bit poetic way, The humanity has navigated the seas for centuries. The artificial domain is still taking shape. We must ensure it receives the same level of attention and care as the maritime security. That is why the President of Latvia was among the world leaders that issued a call for control of frontier AI models earlier this week. And the main premise of the call is very clear. Artificial intelligence must remain under human direction, oversight and control. I think also mentioned by the author of the report. And it must be developed and used in a line with international law. Furthermore, recent Security Council discussion on artificial intelligence, such as the briefing convened by the presidency of France on Wednesday, or informal interactive dialogue organized by Latvia, France, Greece and Bahrain at the beginning of the month, clearly demonstrate the value of Council's engagement. So, as a Security Council member, Latvia is strongly engaged and will be engaged. We are committed to advancing further steps, such as the establishment of councils and a formal expert council on technology. That has also been highlighted in the report that was presented today. We are looking forward to hard work. Thank you very much.

General. And now I will give the floor to the Permanent Representative of the Kingdom of Saudi Arabia. Ambassador, you have the floor.

Thank you, Ambassador. I'd like to join others in expressing my sincere gratitude to the Co-Chairs for organizing this important meeting. The world today stands at a pivotal stage in the course of technological development, as artificial intelligence represents a point of convergence between promising opportunities and significant challenges. It is capable of contributing to the promotion of international peace and security when used responsibly, and it may become a source of tension and instability if its implications are neglected and its technological and ethical complexities are disregarded. From this standpoint, the Kingdom affirms that the responsible and safe use of this development is a fundamental commitment across various fields, including the civilian, military, health, economic, educational, and defense sectors. Aligned with Saudi Vision 2030, which has made artificial intelligence a strategic pillar for accelerating development and empowering human capabilities, the Kingdom has designated 2026 as the year of artificial intelligence, directing its efforts toward improving productivity and public services in health, education, energy, and industrial sectors. The Saudi Data and Artificial Intelligence Authority leads these efforts, supported by quantitative investment in infrastructure and computing capabilities, including the Humane Project. As countries around the world move towards adopting these technologies, ensuring the safety of our citizens from the risks should be taken into consideration in the absence of binding regulatory frameworks that ensure their use in a manner consistent with human values, privacy, and security. Recognizing these complexities, from cybersecurity to the impact of automation on labor markets and environmental sustainability, the Kingdom has adopted a comprehensive approach through national legislation and effective governance frameworks based on transparency and accountability, and combining the artificial intelligence ethics, data protection, and risk management. The Kingdom also contributes to the international governance efforts through the United Nations and Global Partnership on Artificial Intelligence, through its cooperation with UNESCO via the International Center of AI Research and Ethics in Riyadh, in addition to hosting the Global AI Summit in Riyadh from 15 to 17 September 2026, as well as its cooperation with the Islamic World education, scientific and cultural organization in support of the voices of developing countries in a global governance. In conclusion, maintaining international peace and security in the age of artificial intelligence requires sincere cooperation and a partnership that respects human safety, security and rights. The Kingdom renews its commitment to being an active partner in these efforts so that these technologies remain a tool for peace and development and not a means of division and threat. Thank you.

Ambassador, it appears that we have the last speaker before our closing remarks. And I will give the floor to the Deputy Permanent Representative of Armenia for his comments. You have the floor.

Thank you very much, Ambassador Balta. Excellencies, ladies and gentlemen, we would like to express our gratitude to dear friends of Armenia, our co-chairs, Mr. Gheorghe Petritis and Josh Shankar, and their teams, for convening this timely discussion. Artificial intelligence is no longer simply an emerging technology. It is increasingly embedded in our economies, communications, and security environment. The question before us is therefore not whether AI will affect international peace and security, but how we ensure that its development and use contribute to peace, stability, and human development. AI is neither inherently a force for peace nor a source of conflict. Its impact will depend on how we develop, govern, and use it. Our task should be to ensure that AI remains at the service of people, strengthens human capabilities rather than replaces them, and creates new opportunities for human development, economic growth, and cooperation among nations. Armenia is committed to playing its role in shaping and governing the digital era. This commitment is reflected in Armenia's investment in its own AI capabilities under a five-year agreement with Firebird AI. The government of Armenia will procure high performance computing resources to support startups, research groups, academic institutions and individual specialists in the development, training and deployment of AI systems. Distinguished colleagues, we would increasingly explore. We should increasingly explore the practical application of AI in support of peace and security. AI can strengthen conflict early warning by identifying relevant patterns of processing large quantities of information. It can support humanitarian needs assessments, satellite analysis, humanitarian logistics, and reconstruction planning. It can also contribute to ceasefire monitoring and verification through the processing of imagery, sensor data, and other information. These are areas where AI can complement existing capabilities and allow information to be processed more rapidly and effectively. At the same time, its use, particularly in sensitive security contexts, should remain subject to human oversight and responsible governance. Many aspects of our lives could increasingly be delegated to AI, but technology cannot replace diplomacy, political negotiation, trust building, legitimate processes of political domain and accountability, and more importantly, people-to-people contacts. We should hence focus on using AI where it can strengthen our capacity to prevent conflict, respond to humanitarian needs, and support peace while keeping decision making and responsibility firmly in human hands. Thank you.

Thank you, Deputy Permanent Representative, for your comments. And now I don't see anyone. seeking the floor. So I will give the floor to the Deputy Minister of Digital Governance and Artificial Intelligence of Greece, Christos Dermajopoulos, for his closing remarks. Minister, you have the floor.

Thank you, Ambassador. Excellencies, dear foreign ministers, dear undersecretaries, distinguished colleagues, ladies and gentlemen. Allow me, as we bring this event to a close, to thank all participants for timely and valuable discussion on AI and international peace and security. I think that our discussion has underlined that AI is reshaping the global security landscape. It creates new opportunities to improve situational awareness, strengthen early warning capabilities, support crisis management, and protect critical infrastructure. Especially in the maritime domain, AI can enhance maritime awareness, identify suspicious activity, improve navigation safety, and help vast maritime areas more effectively than even before. At the same time, AI can also amplify existing risks, including cyber threats, disinformation, and the misuse of autonomous systems. These opportunities and risks must continue to be addressed together. For Greece, AI is a strategic enabler that's already transforming the maritime security environment, strengthening both maritime domain awareness and the protection and resilience of critical maritime infrastructure. Given Greece's operational presence across the Eastern Mediterranean, the wider Mediterranean, and the Middle East, AI enables the fusion and analysis of vast volumes of satellite, radar, automatic identification systems, radio frequencies, and other data to identify anomalous behavior, dark or shadow vessels, illicit activities, and migration flows, supporting earlier responses and ultimately saving lives at the sea. A key conclusion from today's exchange is that these capabilities must be employed in full respect of the UN Charter, of course, UNCLOS, and the established rules of international law and the law of the sea, which provide the common legal framework for all states and all activities at sea. For Greece, responsible use of AI should reinforce, not weaken, freedom of navigation, respect of sovereign rights, and the rule of international law. The maritime dimension has rightly received particular attention. Growing dependence on undersea cables, ports, logistic networks, and other critical maritime infrastructure creates new vulnerabilities. AI can be used for reconnaissance, vulnerability identification, and increasingly sophisticated cyber operations. Yet, the same technology can strengthen resilience through predictive maintenance, anomaly detection, cyber defense, and real-time monitoring of maritime activities. As a country with the world's largest merchant fleet and a strong maritime tradition, Greece views the protection of maritime infrastructure as a strategic priority. This is why Greece's national approach combines innovation with governance. We have established the Special Secretariat for AI and Data Governance, implementing the national law for the EU AI Act, and we are investing in strategic infrastructure, including the Vethalos supercomputer and the Pharos AI factory. These initiatives reflect our commitment to trustworthy, secure, and human-centric AI. As we conclude, the United Nations has a vital role to play in promoting dialogue, exchanging best practices, strengthening capacity building, supporting international cooperation, and ensuring coherence across security, development, human rights, and technology governance. No country can address these challenges alone. The responsible governance of AI requires a multilateral engagement, shared principles, and collective action. Let this be the message with which we close today's event. AI must serve peace, security, resilience, and sustainable development firmly anchored in international law and human values. Thank you.

On behalf of my colleague and dear friend, the Permanent Representative of Vidia, I would like to thank our briefers, the ministers for being present today, and all of you who have attended our sound event. Thank you. See you to the next one.