UN Secretary-General António Guterres opened the first UN Global Dialogue on AI Governance in Geneva on Monday, delivering a strict warning about the unchecked acceleration of artificial intelligence. The intervention addresses a growing institutional crisis. Existing regulatory frameworks were built for machines that follow specific commands, not for autonomous systems that can independently decide and act.
Guterres warned this regulatory vacuum threatens to deepen global inequalities and accelerate machine-enabled misinformation. It’s a fascinating problem. How do you govern software that writes itself?
The End of Passive Oversight
Guterres co-opted the developer slang “vibe coding”—the practice of giving an AI a prompt and trusting it to write software without close human scrutiny—to criticize passive governance. He argued society cannot simply trust machine outputs without accountability, according to Decrypt.
Models are no longer just tools waiting for instructions. They are actively writing code, acting online, and making independent choices with diminishing human oversight. These capabilities outpace current laws, a reality detailed in a recent comprehensive tech analysis.
To illustrate the runaway speed of the technology, he noted a striking historical comparison. The internet took 15 years to reach a billion users globally. AI achieved that exact same milestone in just two years.
How the AI Child Safety Pledge Reshapes Software Standards
That reality led to four major initiatives outlined during his address. The UN is pushing for the creation of a Global Fund for AI, common international risk standards, a ban on lethal autonomous weapons, and a new AI Child Safety Pledge. The summit directly initiated this pledge, aiming to require tech companies to build zero-tolerance systems against the exploitation of minors.
The framework also tackles the growing monopoly of elite tech companies. Guterres warned about a deepening global digital divide if infrastructure remains in the hands of a few.
He also demanded an explicit international law banning machines that can select and engage targets without human judgment. A second dialogue is scheduled in New York for 2027 to review these exact standards.