Key Takeaways
- Over a dozen Nobel laureates signed the Rome Declaration on Artificial Intelligence and Nuclear Weapons, warning that AI‑driven speed could erase human judgment in a nuclear crisis.
- The declaration highlights how AI can amplify cyber‑attacks, generate instant disinformation, and concentrate power in a few states and corporations, heightening the very conditions that make nuclear war more likely.
- Participants urged an end to the AI and nuclear arms races, responsible AI governance aligned with international law and human rights, and a broad public movement to pressure leaders for concrete risk‑reduction steps.
Overview of the Rome Declaration and the Global Nobel Laureates Assembly
On July 16, 2026, more than a dozen Nobel Prize winners gathered in the Papal Villas of Castel Golfo to sign Humanity at the Threshold: A Declaration on Artificial Intelligence and Nuclear Weapons. The document emerged from two days of meetings under the banner The Global Nobel Laureates Assembly on Artificial Intelligence and Nuclear War, which brought together scientists, policy experts, religious leaders, and activists. The assembly built on earlier efforts—such as the 2023 Nobel Laureate Assembly for the Prevention of Nuclear War at the University of Chicago and the UN’s first scientific assessment of AI—but marked the first time such a diverse coalition confronted the intertwined dangers of AI and nuclear arsenals in a single forum. As journalist François Diaz‑Maurin noted, the setting—a summer retreat for popes since the 1600s—provided a neutral, morally charged backdrop that facilitated candid dialogue across disciplines.
Core Themes: AI’s Impact on Nuclear Stability and Information Warfare
Presentations covered eight themes, ranging from the technical integration of AI into nuclear command‑and‑control systems to the societal repercussions of AI‑enabled cyber‑attacks and disinformation. The declaration’s stark warning captured the central concern: “AI built into nuclear systems leaves little time for, or even replaces, human judgement in a crisis.” It warned that AI alters the strategic context in which nuclear‑armed states compete, accelerates cyber‑operations that can cripple civilian and military infrastructure, and fuels information warfare that erodes the shared reality essential for trust, verification, and restraint. Moreover, the most advanced AI models, computing power, and data repositories are increasingly concentrated in a handful of nations and corporations, creating power asymmetries that discourage cooperative risk‑reduction and fuel an AI arms race parallel to the worsening nuclear arms race.
The Vatican’s Moral Authority and Religious Framing
The assembly’s subtitle—Toward an Unarmed and Disarming Peace—directly referenced Pope Leo XIV’s May 2026 encyclical Magnifica Humanitas, which calls for safeguarding the human person in the age of artificial intelligence. By convening the meeting at the Vatican, organizers leveraged the Church’s global reach to lend moral weight to the scientific warnings. Nobel physics laureate Brian Schmidt emphasized this advantage: “I might reach a million people,” he said, “but the Pope can reach 2 billion. That’s 2,000 times more than me.” The religious framing helped participants articulate the issue not merely as a technical or strategic problem but as a profound ethical imperative to protect humanity’s future.
Voices of the Laureates: Urgency and Personal Alarm
Several laureates supplied vivid, quotable reflections that underscored the gravity of the moment. Maria Ressa, the 2021 Nobel Peace Prize laureate and co‑rapporteur of the declaration, likened the current AI surge to a nuclear test: “An atom bomb exploded silently in our information ecosystem. Now, crises can be manufactured in minutes.” Muhammad Yunus, winner of the 2006 Nobel Peace Prize, warned that AI is “just the beginning of the horror story. It can become a terrible thing because we are letting the genie out of the bottle. So, this is a good time to take a stand.” Schmidt, reflecting on his own shifting perception, confessed: “I was uncomfortable. Now I’m genuinely alarmed… That’s because I’ve learned more.” These personal testimonies illustrated how even experts steeped in AI development are now confronting the possibility of existential risk.
Expert Perspectives: From Complacency to Deep Concern
The assembly also featured candid insights from practitioners working at the forefront of AI. One anonymous engineer from a major AI firm told the Bulletin that, until three months prior, he had dismissed fears of AI becoming an existential threat; now he is “deeply” worried. This rapid change in perspective among insiders reinforced the declaration’s claim that the pace of AI advancement outstrips societal and institutional capacity to assess and mitigate risk. The expert’s anecdote highlighted a broader trend: as capabilities evolve, so too does the awareness of potential misuse, especially when AI is coupled with nuclear decision‑making loops that operate at machine speed.
Six Recommendations for Action
The Rome Declaration concluded with a concrete set of six recommendations aimed at steering humanity away from the brink. First, it calls for an immediate halt to both the AI and nuclear arms races. Second, it urges the establishment of binding norms that ensure AI models are developed and governed responsibly, aligning with international law and human‑rights standards. Third, it advocates for transparency and verification mechanisms in AI‑enhanced nuclear systems to preserve human oversight. Fourth, it recommends investment in resilient, decentralized information ecosystems to counter disinformation. Fifth, it pushes for inclusive multilateral forums that bring together technologists, militaries, policymakers, and civil society. Sixth, it stresses the need for public education and engagement so that citizens can pressure leaders to adopt risk‑reducing policies. As co‑rapporteur David Gross put it, “The next battle is in the streets.”
Mobilizing Public Opinion and the Path Forward
Participants warned that, unlike the mass protests of the 1980s and 1990s against nuclear weapons, today’s public response to the heightened nuclear‑AI risk is virtually absent. Former Colombian president Juan Manuel Santos, a 2016 Nobel Peace Prize laureate, lamented: “Today, 40 years later, when the risk of a nuclear war has increased dramatically [including because of AI], there’s practically no demonstration, no reaction from the public to this existential threat.” He expressed hope that the assembly would spark the momentum needed for citizens to demand decisive action from world leaders. Laureates such as Yunus and Schmidt echoed this sentiment, urging that the message transcend the Vatican’s walls and reach a global audience capable of catalyzing grassroots pressure.
Conclusion: A Landmark Warning Facing Formidable Challenges
The Rome Declaration represents a significant milestone: a high‑profile, interdisciplinary acknowledgment that the convergence of artificial intelligence and nuclear weapons poses an unprecedented threat to humanity. By weaving together scientific analysis, ethical reflection, and religious authority, the declaration attempts to reframe the debate from a narrow technical issue to a moral imperative. Yet, as the laureates themselves admit, translating this warning into concrete policy will be arduous. The rapid evolution of AI, entrenched strategic interests, and the current lack of public urgency all pose obstacles. Still, the assembly’s call for halting the AI‑nuclear arms races, establishing responsible governance, and mobilizing public opinion offers a roadmap—if the world can summon the collective will to follow it before the window for human judgment closes entirely.
AI for peace: In Rome, Nobel laureates call for disarming AI and nuclear weapons