Scientists urge UN countries to limit dangerous uses of AI.

Over 200 prominent figures, including Nobel laureates and researchers at AI giants such as Anthropic, Google, Microsoft, and OpenAI, have signed the "Global Call for AI Red Lines" document addressed to countries attending the United Nations General Assembly. They call for the rapid establishment, by 2026, of "red lines" that cannot be crossed for AI and its dangerous uses.

"Artificial intelligence has immense potential to improve human well-being, but its current trajectory presents unprecedented dangers," reads the document, co-signed by Geoffrey Hinton and Yoshua Bengio, considered among the fathers of AI. "Governments must act decisively before the window for meaningful intervention closes." According to the authors of the letter, the 'red lines' for AI would be internationally agreed-upon bans on uses deemed too risky, including relying on AI to command nuclear arsenals or any type of lethal autonomous weapons system. Or the use of AI for mass surveillance, cyberattacks, and deepfakes.

The signatories of the message urged governments to establish limits on AI by the end of next year, given the pace of technological advancement. "Artificial intelligence could soon far surpass human capabilities and increase risks such as engineered pandemics, widespread disinformation, large-scale manipulation of individuals, including children, national and international security challenges, mass unemployment, and systematic human rights violations," the letter reads. "If nothing is done, many experts, including those on the front lines of development, warn that it will become increasingly difficult to exercise meaningful human control in the coming years."