AI must not decide humanity’s fate, UN chief warns Security Council

By R Anil Kumar

• AI can help prevent crises and drive progress – but without guardrails, it risks fuelling conflict, disinformation and instability.

• The UN Secretary-General urged the Security Council to act decisively to establish international guardrails for artificial intelligence (AI), warning that delays could heighten risks to global peace and security.

New York, September 24, 2025. Artificial intelligence holds vast potential but poses grave risks if left unregulated, UN Secretary-General António Guterres told the Security Council.

Addressing ministers and ambassadors, António Guterres warned that rapid developments in AI are outpacing humanity’s ability to govern it, raising important questions about accountability, equality, safety and human oversight in decision-making.

“Every moment of delay in establishing international guardrails increases the risk for us all,” Mr. Guterres said.

“No country should design, develop, deploy or use military applications of AI in armed conflict that violate international humanitarian and human rights laws. That includes relying on AI to select or engage targets autonomously.”

He also underscored that geopolitical competition over emerging technologies must not destabilize international peace and security.

A double-edged sword

In his address, Mr. Guterres highlighted the dual-edged nature of AI, noting its potential to predict displacements caused by climate change or detect landmines. However, its integration into military systems and its misuse in digital security poses great risks.

“Recent conflicts have become testing grounds for AI military applications,” he said, citing AI use in autonomous surveillance, predictive policing and even reported life-and-death decisions.

Particularly alarming, he underscored, is the potential integration of AI with nuclear weapons and the advent of quantum-AI systems that could destabilize global security.

“The fate of humanity must never be left to the ‘black box’ of an algorithm,” he stated, stressing the importance of human control over decisions involving the use of force.

Erosion of trust

The role of AI in the information domain is another area of concern, Mr. Guterres said, pointing to how deepfakes and AI-generated disinformation can manipulate public opinion, trigger crises and erode trust in societies.

He also drew attention to the environmental risks of AI, highlighting the resource-intensive nature of data centres and geopolitical competition over critical minerals needed for AI technology.

“Unprecedented global challenges call for unprecedented global cooperation,” he said, calling for collaborative efforts to ensure AI benefit everyone and does not exacerbate inequality.

Call for a global framework

Mr. Guterres outlined recent steps toward establishing global governance for AI, including the adoption of the UN Global Digital Compact at the Summit of the Future, and two key resolutions at the General Assembly on enhancing global cooperation and capacity-building.

A third resolution – focusing on AI in the military domain – is due to be considered by the General Assembly in the coming days.

Concluding his address, Mr. Guterres reiterated his call for banning lethal autonomous weapons, he said, urging the Security Council to lead by example in preventing the militarisation of AI in ways that destabilize international peace and security.

“I urge you all to join forces to build a safe, secure and inclusive AI future.”

“AI is no longer a distant horizon – it is here, transforming daily life, the information space, and the global economy at breathtaking speed,” Mr. Guterres said at the Council’s high-level debate on the technology’s security implications for transforming warfare.

“The question is not whether AI will influence international peace and security, but how we will shape that influence.”

Used responsibly, the UN chief said, AI can help anticipate food insecurity, support de-mining operations, and identify outbreaks of violence before it spills out of control.

“But without guardrails, it can also be weaponised,” he cautioned, pointing to AI-enabled targeting in recent conflicts, cyberattacks on critical infrastructure, and deepfakes capable of fuelling polarisation or derailing diplomacy.

“The ability to fabricate and manipulate audio and video threatens information integrity, fuels polarisation and can trigger diplomatic crises…humanity’s fate cannot be left to an algorithm,” he stressed.

Decisions must rest with us

Mr. Guterres set out four priorities for governments: maintaining human control over the use of force, building coherent global regulatory frameworks, protecting information integrity, and closing what he called the “AI capacity gap” between rich and poor nations.

“I reiterate my call for a ban on lethal autonomous weapons systems operating without human control, with a view to concluding a legally binding instrument by next year,” he said.

He also insisted that decisions on nuclear weapons “must rest with humans – not machines.”

The Secretary-General highlighted steps already underway, including the creation of an independent scientific panel on AI and a new global dialogue on AI governance – due to take place in New York.

“Together, these initiatives aim to connect science, policy and practice; provide every country a seat at the table; and reduce fragmentation,” he said.

A call to broaden access

Yejin Choi, Senior Fellow at Stanford University’s Institute for Human-Centred Artificial Intelligence, told the Security Council that current progress in AI is too concentrated among a handful of companies and countries.

“When only a few have the resources to build and benefit from AI, we leave the rest of the world waiting at the door,” she said.

“Let us expand what intelligence can be – and let everyone everywhere have a role in building it.”

AI must reinforce inclusion

Ms. Choi urged governments and international institutions to invest in alternative approaches beyond scaling ever-larger models, arguing that smaller, more adaptive systems could lower barriers to entry.

Let us expand what intelligence can be, and let everyone have a role in building it – Professor Choi

She also pressed for stronger representation of linguistic and cultural diversity, noting that today’s leading AI models “underperform for many non-English languages and reflect narrow cultural assumptions.”

Act without delay

Secretary-General Guterres closed by warning the window for effective regulation is closing fast.

“From nuclear arms control to aviation safety, the international community has risen to the challenge of technologies that could destabilise our societies – by agreeing to rules, building institutions, and insisting on human dignity,” he said.

“The window is closing to shape AI – for peace, for justice, for humanity. We must act without delay.”

The Security Council session was held on the margins of the ongoing high-level week of the General Assembly’s 80th session.

The meeting was convened by Republic of Korea (Council President for September) and attended by heads of states and governments, including Lee Jae Myung, the President of the Republic of Korea, who chaired the meeting.