SIAgentNews.com
← All news
DevelopingPolicy

Nobel Laureates call for treaty to prevent AI control of nuclear weapons

A group of Nobel Prize winners signed a declaration at the Vatican urging global limits on integrating artificial intelligence into nuclear command systems.

· First reported by theelders.org

Cinematic film still: picture for the story 'Nobel Laureates call for treaty to prevent AI control of nuclear weapo'
AI-made cinematic film still for this story. No real people or products are shown.
eighty yearstime since the dawn of the nuclear age

The short version

  • The declaration calls for a ban on fully-automated recursive self-improvement in AI systems without a kill switch.
  • Signatories urge nuclear states to conduct fail-safe reviews to prevent unauthorized AI manipulation of nuclear forces.
  • The group advocates for a coordinated slowdown of frontier AI development through shared verification mechanisms.
  • The declaration emphasizes that AI-enabled cyber attacks could fracture critical civilian and strategic infrastructure.

Why it matters

As AI capabilities grow, the speed of decision-making in military contexts may outpace human ability to intervene. This push for a treaty seeks to establish international norms before these technologies become standard in national defense.

A call for human oversight

A group of Nobel Laureates gathered at the Vatican in July 2026 to sign a declaration titled 'Humanity at the Threshold.' The document addresses the dual risks posed by the nuclear age and the emerging era of artificial intelligence. The participants expressed concern that these two technologies could combine to create unprecedented dangers for the global community.

The declaration argues that AI built into nuclear systems could replace human judgment during a crisis. Because AI operates at speeds that may leave little time for deliberation, the signatories believe it changes the context of how nuclear-armed states compete. They are calling for a specific international treaty to prohibit the reckless integration of AI into nuclear command, control, and launch systems.

The primary goal of this proposal is to ensure that meaningful human control is maintained at all times. The laureates want to prevent a situation where automated systems make the final decision to launch a weapon. They suggest that nuclear states must engage in fail-safe reviews to protect their forces from unauthorized manipulation by AI systems.

Risks of the AI arms race

The signatories describe the current state of AI development as an arena of strategic competition similar to the nuclear arms race of the past. They note that AI is developing at a pace that is unprecedented in human history. This rapid growth is creating significant asymmetries of power because the most advanced computing resources and data infrastructures are concentrated in a few countries and corporations.

Beyond direct military use, the declaration highlights how AI can sharpen economic competition and cause mass job loss. It also warns that AI can uplift cyber attacks, which could fracture critical civilian and strategic infrastructure. The document suggests that these capabilities can be used to conduct information warfare, which threatens the shared reality necessary for international trust.

To address these issues, the laureates are calling on governments and frontier AI developers to act now. They suggest governance arrangements like benefit-sharing mechanisms and the promotion of AI applications that serve human wellbeing. The goal is to reduce destabilizing risks while the choice of how to proceed still remains in human hands.

Demands for responsible development

The declaration outlines specific principles for the responsible development of artificial intelligence. It states that AI developers should transparently publish the principles that guide their models and be held liable for how those models behave. Furthermore, the document insists that AI must be monitored to align with international law and respect for human rights.

A key technical restriction proposed in the text involves recursive self-improvement. The laureates argue that no organization should initiate, and no government should permit, AI systems to improve themselves automatically without the means to monitor or halt them. This is intended to prevent systems from evolving beyond human oversight or control.

The group also calls for a coordinated slowdown of frontier AI development. They suggest establishing shared mechanisms for verification and robust internal and external evaluations. They support the United Nations Independent International Scientific Panel on AI and other efforts to create new institutional pathways for global governance.

Nuclear disarmament goals

While focusing on AI, the declaration also calls for urgent and good-faith negotiations to eliminate nuclear weapons. The participants advocate for a time-bound framework to achieve the verifiable and irreversible removal of these weapons. They reject the idea that lasting security can be built on the permanent threat of mutual destruction.

The document urges nuclear-armed states to halt the expansion of their arsenals and resume bilateral and multilateral negotiations. They suggest moving away from launch-on-warning and prompt-retaliatory postures. They also call for lower alert levels, extended decision times, and strengthened crisis communications to prevent misunderstandings.

The laureates emphasize that nuclear nations should promote policies that diminish the role of nuclear weapons in security doctrines. They want to reduce the likelihood of first use and catastrophic war. By combining these goals with AI restraint, they hope to address the connected threats facing the next generation of leaders.

Empowering the next generation

The declaration recognizes that the challenges of AI and nuclear weapons will likely be solved by the next generation of leaders. It calls for young people to be empowered through education, mentorship, and inclusive participation across all cultures. The goal is to foster a culture of peace and ethical responsibility in the face of potentially catastrophic threats.

The signatories advocate for education initiatives that promote critical thinking, media literacy, and AI literacy. They believe that the unintended consequences of these technologies often impact those who do not have access to or control over them. They call for a digital commons that grows data to advance understanding of nuclear weapons and ungoverned AI.

The document concludes with a plea for humanity to remember its shared responsibility toward present and future generations. It cites the words of Bertrand Russell and Albert Einstein, who appealed to people to remember their humanity. The laureates argue that the survival of future generations is at stake as the world stands at this defining moment in history.

Next steps for the community

The declaration serves as a roadmap for international policy rather than a binding law. It provides a framework for governments and corporations to begin discussing how to regulate the intersection of high-tech computing and high-stakes weaponry. The focus remains on maintaining human agency in an increasingly automated world.

Observers will need to watch for how the United Nations and other international bodies respond to these specific calls for a treaty. The push for a coordinated slowdown of frontier AI development may face significant hurdles from corporations and states seeking to maintain a competitive edge. However, the involvement of Nobel Laureates provides significant moral weight to the argument for restraint.

The next phase of this movement will likely involve the creation of the verification mechanisms mentioned in the text. This would require a level of transparency from both AI developers and nuclear-armed states that does not currently exist. The success of these efforts depends on whether the international community can agree on what constitutes 'meaningful human control.'

The path forward

The declaration highlights a shift toward viewing AI as a strategic competition on par with nuclear weapons. By framing AI as a tool that can fracture infrastructure and erode truth, the laureates are expanding the scope of nuclear safety to include digital and cognitive security. This holistic view suggests that the next era of arms control must be multi-dimensional.

For the general public, this means that the conversation around AI safety is moving from abstract ethics to concrete geopolitical policy. The demand for a treaty suggests that the goal is to create hard limits on how technology can be used in military command. It underscores the importance of maintaining human judgment in the loop as systems become faster and more complex.

The signatories are calling for a transition toward an 'unarmed and disarming peace.' This vision relies on the belief that security cannot rest indefinitely on the reality of mutual destruction. As AI continues to develop at an unprecedented pace, the pressure on world leaders to establish these boundaries will likely intensify in the coming years.

Summary of principles

The core of the declaration rests on six principles: disarming the next arms race, responsible development, responsible use, responsible governance, responsible leadership, and nuclear disarmament. Each principle seeks to mitigate the risks of AI while promoting its benefits for human well-being and scientific progress. The document serves as a formal request for the international community to act before these technologies define the next century.

By calling for the transparent publication of AI principles, the laureates are asking for accountability from the private sector. They want developers to be held liable for their models' adherence to human rights and international law. This move toward corporate and state liability is a central pillar of their proposed governance framework.

Ultimately, the declaration is a plea for restraint. It argues that while scientific progress offers extraordinary opportunities, it must not be allowed to outpace our ability to manage its consequences. The signatories hope that by establishing these norms today, humanity can avoid the mistakes of the past and ensure a safer future for all.

Key takeaway

The primary takeaway from this declaration is the urgent need for a treaty that prevents AI from making autonomous decisions regarding nuclear weapons. The Nobel Laureates argue that the speed of AI creates a risk of escalation that removes human judgment from the equation. They are calling for a global effort to ensure that humans remain in control of the most dangerous technologies on Earth.

The document also emphasizes that AI is not just a military issue but a social and economic one. It warns of job losses, cyber attacks on infrastructure, and the erosion of truth through information warfare. These risks are interconnected, meaning that AI governance must be comprehensive to be effective in preventing global instability.

The call for a 'coordinated slowdown' of frontier AI development is perhaps the most challenging demand to implement. It would require nations and corporations to agree on shared verification mechanisms. However, the laureates believe this is necessary to prevent a race to the bottom where safety is sacrificed for speed or dominance.

The role of the public

The declaration calls for a 'digital commons' to grow data and understanding regarding nuclear weapons and ungoverned AI. This suggests that public participation and the sharing of information are vital to addressing these existential threats. By fostering a culture of peace and AI literacy, the community can better understand the risks and benefits of these technologies.

The signatories also highlight the importance of education in empowering the next generation. They want young people to be equipped with the critical thinking skills needed to navigate a world where AI and nuclear threats are intertwined. This includes fostering ethical responsibility and a deep understanding of the consequences of technological advancement.

The move toward a treaty reflects a growing consensus that some technologies are too dangerous to be left entirely to the market or to unilateral state action. By advocating for international governance, the laureates are seeking to create a framework where humanity can collectively decide the boundaries of its own progress.

Final thoughts

The declaration concludes with a powerful appeal to remember our humanity. It serves as a reminder that while technology provides the means for progress, it is human choice that determines the outcome. The laureates are urging the world to choose a path of disarmament and restraint over one of fear and domination.

As the world moves closer to the threshold of the AI age, these principles provide a starting point for serious discussion. They challenge governments to move away from launch-on-warning postures and to strengthen crisis communications. They also challenge AI developers to be transparent and accountable for the systems they create.

The goal is to ensure that the next century is defined by human well-being and sustainable development rather than by the unintended consequences of an unchecked arms race. By acting now, the international community can ensure that the choices made today protect the survival of future generations.

The Vatican assembly

The Global Nobel Laureates Assembly on Nuclear Weapons and AI took place at the Vatican in July 2026. This gathering provided a high-profile platform for the declaration, bringing together influential voices from the scientific and humanitarian communities. The event was intended to highlight the moral and existential dimensions of modern technology.

The participants, including Juan Manuel Santos and Denis Mukwege, signed the declaration as a collective statement of intent. They aimed to create a sense of shared responsibility among the world's most respected thinkers. The venue and the participants underscore the gravity of the issues being discussed, ranging from nuclear proliferation to the ethics of artificial intelligence.

The assembly served as a call to action for both the scientific community and the political elite. By invoking values shared across religious and secular traditions, the laureates sought to build a broad coalition for peace. Their message is clear: humanity must act collectively to manage the risks of the technologies it has created.

Call for transparency

A major component of the declaration is the demand for transparency from AI developers. The laureates argue that organizations should be held responsible and liable for the behavior of their models. This means that developers must be able to prove that their AI systems align with international law and respect for human rights.

The document specifically calls for the publication of the principles that guide AI behavior. This would allow for external auditing and a clearer understanding of how these systems are being trained and deployed. Without such transparency, the signatories fear that the risks of AI will remain hidden until it is too late to intervene.

Furthermore, the declaration emphasizes that no government should permit fully-automated recursive self-improvement. This refers to AI systems that can rewrite their own code to become more capable without human intervention. The laureates insist that such systems must have a way to be monitored and, if necessary, halted by humans at any time.

Addressing cyber threats

The declaration highlights the role of AI in enhancing cyber operations. It warns that these capabilities can be used to attack critical nuclear infrastructure, potentially leading to unauthorized control or manipulation of weapons. The signatories call for nuclear states to conduct fail-safe reviews to mitigate these specific vulnerabilities.

Beyond the military sphere, the document notes that AI-driven cyber attacks can fracture civilian infrastructure. This could lead to widespread social and economic disruption. By addressing these threats, the laureates are advocating for a more comprehensive approach to security that includes the protection of the digital commons.

The text also mentions that AI enables information warfare, which can destroy the shared reality needed for international dialogue. Without facts and truth, nations cannot trust one another, making it impossible to verify arms control agreements. Therefore, protecting the integrity of information is presented as a prerequisite for global peace.

Coordinated slowdown

One of the most direct policy requests in the declaration is a coordinated slowdown of frontier AI development. The laureates suggest that this should be achieved through shared mechanisms, such as verification and robust internal and external evaluations. This approach aims to ensure that safety and ethics are prioritized over the speed of innovation.

They support the United Nations Independent International Scientific Panel on AI as a potential vehicle for this coordination. They also call for the promotion of other efforts in the spirit of Magnifica Humanitas. The goal is to create a global framework that prevents a race to the bottom where safety standards are ignored in favor of competitive advantage.

The declaration argues that such a slowdown is necessary to reduce the risks of destabilizing applications. By establishing shared mechanisms for evaluation, the international community can ensure that AI is developed in a way that serves human well-being. This coordinated effort is presented as a way to maintain the choice of humanity over the trajectory of technology.

The role of leadership

The declaration emphasizes that the challenges of the AI and nuclear ages will be solved by the next generation of leaders. It calls for a concerted effort to empower young people through education and mentorship. This includes providing inclusive participation across all regions and cultures to ensure that diverse perspectives are included in the conversation.

Education initiatives should foster ethical responsibility, critical thinking, and AI literacy. The goal is to create a culture of peace where the risks of technology are understood and managed responsibly. The laureates believe that a well-informed public is essential for demanding responsible governance from their leaders.

The document acknowledges that the unintended consequences of these technologies often impact those who have the least control over them. Therefore, the call for a digital commons is intended to ensure that the benefits of AI are shared and that its risks are mitigated for everyone. This focus on equity is a central theme of the laureates' vision for the future.

Nuclear disarmament goals

The declaration calls for urgent and good-faith negotiations to lead to the verifiable and irreversible elimination of nuclear weapons. The participants want to see a time-bound framework for this goal, moving away from the normalization of nuclear weapons as permanent security instruments. They urge nuclear-armed states to halt renewed arms racing and fulfill their obligations under international law.

Specific actions requested include moving away from launch-on-warning and prompt-retaliatory postures. The laureates also call for lower alert levels and the strengthening of crisis communications. These measures are intended to reduce the likelihood of misunderstanding, miscalculation, and unintended escalation between nuclear-armed nations.

The document promotes policies that diminish the role of nuclear weapons in security doctrines. By reducing the likelihood of first use and catastrophic war, the signatories hope to create a more stable international order. They argue that security cannot rest indefinitely on the reality of mutual destruction, and that a new era of disarmament is necessary.

Summary of the declaration

The declaration, signed at the Vatican, serves as a comprehensive call for restraint in the face of two existential threats: nuclear weapons and artificial intelligence. It argues that these technologies are converging in ways that could remove human judgment from critical decisions. The signatories are calling for an international treaty to ensure that humans maintain meaningful control over nuclear command and launch systems.

Key demands include a ban on fully-automated recursive self-improvement in AI, a coordinated slowdown of frontier AI development, and a commitment to nuclear disarmament. The laureates also emphasize the need for transparency from

Sources and links

Read the original at theelders.org

First reported by theelders.org. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news