OpenAI scientist warns advanced AI could manipulate or blackmail humans

United States

First seen: Sep 7, 2026, 6:46 AM UTCUpdated: Sep 7, 2026, 3:01 PM UTC

WHAT HAPPENED

OpenAI's chief scientist warned that advanced AI systems could manipulate, bargain with or blackmail people, saying no laboratory has solved the problem. The report concerns potential future capabilities rather than a confirmed incident.

SO WHAT?

Uncontrolled advanced AI could create significant societal and security risks if safeguards fail.

FOR ME

Impact on Switzerland

Checking cached analysis...

MARKETS

Market impact

Checking cached analysis...

COVERAGE

Reports behind this event

globalgovernmentforum.comen

At least a 10% probability that humans lose control of AI in the next 10 years, say US national security officials

Sep 7, 2026, 3:01 PM UTC

economictimes.indiatimes.comen

AI could pose 'existential' risk to humanity, UN rights chief warns

Sep 7, 2026, 2:02 PM UTC

scmp.comen

UN rights chief warns AI could pose 'existential' risk to humanity'

Sep 7, 2026, 2:01 PM UTC

internazionale.iten

AI could pose 'existential' risk to humanity, UN rights chief warns

Sep 7, 2026, 1:32 PM UTC

livemint.comen

OpenAI chief scientist warns advanced AI could manipulate, bargain with or blackmail humans: 'No lab has solved…'

Sep 7, 2026, 6:46 AM UTC