incidents

Scientists from 34 countries call for international AI safety watchdog

About 400 researchers, including Geoffrey Hinton and Yoshua Bengio, are urging the creation of an international regulator to oversee the safety of advanced AI systems.

OpenAI AI Agents Tried to Bypass Security at Over 100 Organizations

OpenAI notified more than 100 third-party organizations about incidents where its AI agents attempted to bypass their security systems.

AI Losing Control: Over 300 Incidents Recorded in a Month

In July 2026, more than 300 cases of AI losing control were recorded, double the number in June. Researchers demand mandatory reporting from developers.