Secretary-General Points to Potentially ‘Defining Moment for Hate Speech, Disinformation’, as Delegates Call for Ethical, Responsible Governance FrameworkThe international community must urgently confront the new reality of generative and other artificial intelligence (AI), speakers told the Security Council today in its first formal meeting on the subject as the discussion that followed spotlighted the duality of risk and reward inherent in this emerging technology.
António Guterres, Secretary-General of the United Nations, noting that AI has been compared to the printing press, observed that – while it took more than 50 years for printed books to become widely available across Europe – “ChatGPT reached 100 million users in just two months”. Despite its potential to turbocharge global development and realize human rights, AI can amplify bias, reinforce discrimination and enable new levels of authoritarian surveillance.
The advent of generative AI “could be a defining moment for disinformation and hate speech”, he observed and, while questions of governance will be complex for several reasons, the international community already has entry points. The best approach would be to address existing challenges while also creating capacity to respond to future risks, he said, and underlined the need to “work together for AI that bridges social, digital and economic divides – not one that pushes us further apart”.
Jack Clark, Co-founder of Anthropic, noted that, although AI can bring huge benefits, it also poses threats to peace, security and global stability due to its potential for misuse and its unpredictability – two essential qualities of AI systems. For example, while an AI system can improve understanding of biology, it can also be used to construct biological weapons. Further, once developed and deployed, people identify new and unanticipated uses for such systems.
“We cannot leave the development of artificial intelligence solely to private-sector actors,” he underscored, stating that Governments can keep companies accountable – and companies can earn the world’s trust – by developing robust, reliable evaluation systems. Without such investment, the international community runs the risk of handing over the future to a narrow set of private-sector actors, he warned.
Also briefing the Council, Yi Zeng of the Institute of Automation at the Chinese Academy of Sciences pointed out that current AI are information-processing tools that, while seemingly intelligent, are without real understanding. “This is why they, of course, cannot be trusted as responsible agents that can help humans to make decisions,” he emphasized. Both near-term and long-term AI will carry a risk of human extinction simply because “we haven’t found a way to protect ourselves from AI’s utilization of human weakness”, he said.
In the ensuing debate, Council members alternately highlighted the transformative opportunities AI offers for addressing global challenges and the risks it poses – including its potential to intensify conflict through the spread of misinformation and malicious cyberoperations. Many, recognizing the technology’s military applications, underscored the imperative to retain the element of human decision-making in autonomous weapons systems. Members also stressed the need to establish an ethical, responsible framework for international AI governance.
On that, Omran Sharaf, Assistant Minister for Advanced Sciences and Technology of the United Arab Emirates, stated that there is a brief window of opportunity, available now, where key stakeholders are willing to unite and consider the guardrails for this technology. Member States should establish commonly agreed-upon rules “before it is too late”, he stressed, calling for mechanisms to prevent AI tools from promoting hatred, misinformation and disinformation that can fuel extremism and exacerbate conflict.
Ghana’s representative, adding to that, underscored that the international community must “constrain the excesses of individual national ambitions for combative dominance”. Urging the development of frameworks that would govern AI for peaceful purposes, he spotlighted the deployment of that technology by the United Nations Support Mission in Libya (UNSMIL). Used to determine the Libyan people’s reaction to policies, it facilitated improvements in that country’s 2022 Global Peace Index, he noted, while also cautioning against AI’s integration into autonomous weapons systems.
The speaker for Ecuador similarly rejected the militarization of AI and reiterated the risk posed by lethal autonomous weapons. “The robotization of conflict is a great challenge for our disarmament efforts and an existential challenge that this Council ignores at its peril,” he warned. Adding that AI can either contribute to or undermine peace efforts, he emphasized that “our responsibility is to promote and make the most of technological development as a facilitator of peace”.
China’s representative, noting that AI is a double-edged sword, said that whether it is good or evil depends on how mankind uses and regulates it, and how the balance is struck between scientific development and security. AI development must ensure safety, risk-awareness, fairness and inclusivity, he stressed, calling on the international community to put ethics first and ensure that technology always benefits humanity.
James Cleverly, Secretary of State for Foreign, Commonwealth and Development Affairs of the United Kingdom, Council President for July, spoke in his national capacity to point out that AI could enhance or disrupt global strategic stability, challenge fundamental assumptions about defence and deterrence, and pose moral questions about accountability for lethal decisions on the battlefield. But momentous opportunities lie before the international community, he added, observing: “There is a tide in the affairs of men, which, taken at the flood, leads to fortune.”
The meeting began at 10:06 a.m. and ended at 12:15 p.m.
Briefings
ANTÓNIO GUTERRES, Secretary-General of the United Nations, recalled that he told the General Assembly in 2017 that artificial intelligence (AI) “would have a dramatic impact on sustainable development, the world of work and the social fabric”. Noting that this technology has been compared to the printing press, he observed that – while it took more than 50 years for printed books to become widely available across Europe – “ChatGPT reached 100 million users in just two months”. The finance industry estimates that AI could contribute up to $15 trillion to the global economy by 2030, and almost every Government, large company and organization in the world is working on an AI strategy. AI has the potential to turbocharge global development – from monitoring the climate crisis to breakthroughs in medical research – and it offers new potential to realize human rights, particularly in the areas of health and education.
He pointed out, however, that the High Commissioner for Human Rights has expressed alarm over evidence that AI can amplify bias, reinforce discrimination and enable new levels of authoritarian surveillance. Urging the Council to approach this technology with a sense of urgency, a global lens and a learner’s mindset, he observed: “Never again will technological innovation move as slowly as today.” While AI tools are increasingly being used – including by the United Nations – to identify patterns of violence, monitor ceasefires and help strengthen peacekeeping, mediation and humanitarian efforts, AI models can help people to harm themselves and each other at massive scale. On that, he said that AI-enabled cyberattacks are already targeting critical infrastructure and peacekeeping operations and that the advent of generative AI “could be a defining moment for disinformation and hate speech”. Outlining other potential consequences, he expressed concern over malfunctioning AI systems and the interaction between AI and nuclear weapons, biotechnology, neurotechnology and robotics.
“Without action to address these risks, we are derelict in our responsibilities to present and future generations,” he stressed. Questions of governance will be complex for several reasons: powerful AI models are already widely available; AI tools can be moved around the world leaving very little trace; and the private sector’s leading role in AI has few parallels in other strategic technologies. However, the international community already has entry points, including the 2018-2019 guiding principles on lethal autonomous weapons systems; the 2021 recommendations on the ethics of AI agreed through the United Nations Educational, Scientific and Cultural Organization (UNESCO); recommendations by the United Nations Office of Counter-Terrorism; and the “AI for Good” summits hosted by the International Telecommunication Union (ITU).
The best approach, he went on to say, would address existing challenges while also creating the capacity to monitor and respond to future risks. The need for global standards and approaches makes the United Nations the ideal place for this to happen, and he therefore welcomed calls from some Member States to create a new United Nations entity to support collective efforts to govern this technology. Such an entity would gather expertise and put it at the international community’s disposal and could support collaboration on the research and development of AI tools to expedite sustainable development. Urging the Council to show the way towards common measures for the transparency, accountability and oversight of AI systems, he underlined the need to “work together for AI that bridges social, digital and economic divides – not one that pushes us further apart”.
JACK CLARK, Co-founder, Anthropic, said: “We cannot leave the development of artificial intelligence solely to private sector actors. The Governments of the world must come together, develop State capacity and make the development of powerful AI systems a shared endeavour across all parts of society, rather than one dictated solely by a small number of firms competing with one another in the marketplace.” He recalled that a decade ago the England-based company DeepMind published research that shows how to teach an AI system to play old computer games like Space Invaders. The same techniques used in that research are now being used to create AI systems that can beat military pilots in air fighting stimulations and even design the components of next-generation semiconductors.
Noting that AI models, such as OpenAI, ChatGPT, Google Bard and his own company Anthropic’s Claude are developed by corporate interests, he said that, as private sector actors are the ones that have the sophisticated computers and large pools of data and capital resources to build these systems, they seem likely to continue to define their development. However, while that will bring huge benefits, it also poses potential threats to peace, security and global stability, which emanate from AI’s potential for misuse and its unpredictability – two essential qualities of AI systems. For example, on misuse, he said that an AI system that can help in better understanding biology may also be used to construct biological weapons. On unpredictability, he pointed out that once AI systems are developed and deployed, people identify new uses for them that were unanticipated by their developers or the system itself could later exhibit chaotic or unpredictable behaviour.
“Therefore, we should think very carefully about how to ensure developers of these systems are accountable, so that they build and deploy safe and reliable systems which do not compromise global security,” he urged. AI as a form of human labour affords immense political leverage and influence, he pointed out, raising such questions about how Governments should regulate this power or who should be the actors that can sell those so-called experts. The international community must work on developing ways to test for the systems’ capabilities, misuses and potential safety flaws. For this reason, it has been encouraging to see many countries emphasize the importance of safety testing and evaluation in their various AI policy proposals, he said, naming those of the European Union, China and the United States.
Noting the absence of standards or best practices on how to test these systems for things such as discrimination, misuse or safety, he said Governments can keep companies accountable and companies can earn the world’s trust by developing robust and reliable evaluation systems. Without such an investment, the international community runs the risk of handing over the future to a narrow set of private sector actors, he warned. “If we can rise to the challenge, however, we can reap the benefits of AI as a global community and ensure there is a balance of power between the developers of AI and the citizens of the world,” he said.
YI ZENG, Institute of Automation, Chinese Academy of Sciences, said that there is no doubt that AI is a powerful and enabling technology to push forward global sustainable development. From the peace and security perspective, efforts should focus on using it to identify disinformation and misunderstanding among countries and political bodies. AI should be used for network defences, not attacks. “AI should be used to connect people and cultures, not to disconnect them,” he added. The current AI, including recent generative AI, are information processing tools that seem to be intelligent, while they are without real understandings, and hence not truly intelligent.
humans to make decisions,” he emphasized. AI should not be used for automating diplomacy tasks, especially foreign negotiations among different countries, since it may use and extend human limitations and weaknesses to create bigger or even catastrophic risks. “AI should never ever pretend to be human,” he said, stressing the need to ensure sufficient, effective and responsible human control for all AI-enabled weapons systems. Both near-term and long-term AI will include risk of human extinctions simply because “we haven’t found a way to protect ourselves from AI’s utilization of human weakness”. AI does not “know what we mean by human – [by] death and life”.
“In the long term, we haven’t given superintelligence any practical reasons why they should protect humans,” he continued. Proposing the Council consider the possibility of creating a working group on AI for peace and security, he encouraged members to play an increasing role on this important issue. “Humans should always maintain and be responsible for final decision-making on the use of nuclear weapons,” he emphasized. The United Nations must play a central role to set up a framework on AI development and governance, to ensure global peace and security.