European lawmakers, Nobel laureates, ex-leaders, and AI researchers launched a campaign at the UN General Assembly.
They urged governments to establish binding “red lines” by 2026 to ban the most dangerous AI applications.
Supporters warned unchecked systems could spark pandemics, disinformation, mass surveillance, and catastrophic rights violations.
Risks Highlight Urgent Need for Safeguards
A new psychiatric study found leading chatbots gave inconsistent or unsafe advice about suicide.
Researchers warned these failures could intensify mental health crises and even contribute to deaths.
Maria Ressa cautioned AI could create “epistemic chaos,” while Yoshua Bengio said societies cannot manage model races.
Push for a Global Treaty by 2026
Over 200 figures and 70 organisations signed the appeal, including OpenAI, Google DeepMind, and Anthropic leaders.
Backers proposed bans on AI-driven nuclear strikes, human impersonation, and surveillance.
They called for an independent oversight body and hope the UN begins treaty negotiations by 2026.
