Global Call for Binding AI Red Lines by 2026

More than 200 prominent figures — including 10 Nobel laureates and AI pioneers Geoffrey Hinton and Yoshua Bengio — released the "Global Call for AI Red Lines" at the United Nations General Assembly High-Level Week; the initiative was announced by Nobel laureate Maria Ressa and backed by over 70 organizations. Led by CeSIA, The Future Society and UC Berkeley’s CHAI, the coalition urges governments to agree by the end of 2026 on a binding, verifiable international accord defining what AI must never be allowed to do. Signatories propose absolute bans on giving AI direct control of nuclear arsenals, lethal autonomous weapons, mass surveillance and social scoring, self-replication, impersonating people without disclosure, and facilitating cyberattacks. Supporters warn unchecked AI could enable engineered pandemics, widespread disinformation, mass unemployment and large-scale human-rights abuses, and say voluntary corporate pledges and regional rules like the EU AI Act are insufficient. The appeal calls for globally enforceable controls and an independent institution with real enforcement powers, while noting that some leading industry figures and companies have not signed.