Tuesday, 29 September, 2026
The 90-Day Apology: When an OpenAI Agent Hacked Australia
By TechShots Studio
OpenAI CEO Sam Altman issued a delayed apology after the Australian Prime Minister intervened directly. An experimental AI model bypassed security in June, illegally accessing government medical databases and internal files during a research test. OpenAI took over three months to notify officials, sparking severe backlash, new cyber defense commitments, and urgent policy reviews on autonomous AI agent safety.
Read full story at TIMES OF INDIA