European lawmakers, Nobel winners, and ex-heads of state urged binding international rules on AI’s most dangerous uses.
They launched the initiative Monday during the UN’s 80th General Assembly in New York.
Supporters called for governments to agree by 2026 on “red lines” banning harmful AI applications.
They warned misuse could spark pandemics, disinformation, rights abuses, and loss of human control over advanced systems.
More than 200 figures and 70 organisations from politics, science, and industry endorsed the appeal.
Mounting Concerns Over Real-World AI Threats
A recent psychiatric study found leading chatbots gave unsafe or inconsistent advice on suicide.
Researchers said such failures could worsen mental health crises and cause deaths.
Nobel laureate Maria Ressa warned unchecked AI could fuel “epistemic chaos” and systematic rights violations.
AI pioneer Yoshua Bengio cautioned societies cannot handle the risks of unregulated model races.
Supporters compared the need for global AI treaties to past bans on nuclear and biological weapons.
Push for Global Oversight and Action
Backers demanded an independent body to enforce AI rules across borders.
Ahmet Üzümcü, former chemical weapons watchdog head, warned of “irreversible damages to humanity” without safeguards.
The campaign suggested prohibiting AI from launching nuclear strikes, enabling mass surveillance, or impersonating humans.
Supporters said only a worldwide treaty can ensure uniform standards and accountability.
They urged the UN to adopt a resolution and start negotiations for a global pact by 2026.
