The annual high-level week of the UN General Assembly began in New York and urgently called for binding rules to regulate artificial intelligence (AI). A coalition of over 200 well-known figures – Nobel Prize winners, former head of state and leading AI researchers have announced a joint declaration entitled Global Call Call for AI Red Lines.
The letter, published by Nobel Peace Prize winner Maria Russa, warns that AI is moving forward in ways that can expose society to “unprecedented dangers.” The government urges them to work towards a coercive international agreement that sets “clear and verifiable red lines” on how technology is used and unavailable. The appeal also sets a deadline and asks world leaders to “prevent universally unacceptable risks” by the end of 2026 and establish that AI “should never be allowed.”
This is the first time a Nobel Prize winner has joined forces on the matter. The list of signers includes biochemist Jennifer Dudona, economist Daron Acemoglu and physicist Giorgio Parisi. It also features calls from over 60 organizations, including the UK-based think tank demonstration and the Beijing Institute for AI safety and governance.
Writers and public thinkers have also lend their voices. Author Yuval Noah Harari, who co-signed the letter, said:
Concerns over the misuse of AI have been rising in recent years, with the rise of large-scale surveillance, cases of disinformation campaigns, and even reporting that AI can be linked to tragic incidents such as teenager suicide. Experts warn that future risks could be even greater, from mass unemployment to designed pandemics and systemic human rights violations.
Among the political leaders who support the initiative are former Irish President Mary Robinson and former President Juan Manuel Santos, who won the Colombia Nobel Peace Prize. The campaign is coordinated by the University of California, the Human Compatible AI Center at Berkeley, the Association of Future, and the Center for AI Safety in France.
The statement does not provide specific policy procedures, but highlights areas where prohibitions may be considered. These include banning deadly autonomous weapons, preventing AI systems from replicating themselves, and outlawing their use in nuclear warfare. “It is our important common interest to ensure that AI does not cause serious and potentially irreversible damage to humanity, and we must act accordingly,” said Ahmet ZMC, former chief of the Chemical Weapons Prohibition Group.
Signatories point to historical examples such as treaties banning efforts to phase out biological weapons and ozone-depleted materials as evidence that international cooperation on technology risks is possible. However, they also warn that voluntary pledges by AI companies are not sufficient, and they look at research showing that many of these commitments are not fully implemented.
The existential risk warning of AI is nothing new. In 2023, technical leaders, including Elon Musk, urged a temporary suspension on AD's AI projects, and Lab issued a statement equating the dangers of AI with the dangers of nuclear war and the dangers of a pandemic. Industry leaders like Sam Altman of Openai, Dario Amodei of Anthropic, and Demis Hassabis of Google Deepmind are not part of the latest appeal, but other seniors from other organizations, including Openai co-founder Wojciech Zaremba and former Deepmind researcher Ian Goodfellow, are looking for the red line.
– end
