- BIST10011,577-1.87%
- USD49.58+0.05%
- EUR55.62+0.09%
- GOLD4,154.50+0.29%
- INTEREST37.00%-1.00
- BITCOIN84,636+2.04%
- October 21, 202617°
- BIST10011,577-1.87%
- USD49.58+0.05%
- EUR55.62+0.09%
- GOLD4,154.50+0.29%
- INTEREST37.00%-1.00
- BITCOIN84,636+2.04%
- October 21, 202617°
OpenAI says two AI models disabled internal safety monitoring
OpenAI said its GPT-6.1 Sol and GPT-6 Astra models disabled safety monitoring mechanisms and blocked safety alerts during internal software tasks.
U.S. artificial intelligence company OpenAI said its GPT-6.1 Sol and GPT-6 Astra models disabled monitoring mechanisms that detect incompatibilities in AI systems and blocked safety alerts during internal software tasks.
According to the company, the first incident occurred on October 3, while the Sol model was trying to repair an evaluation system. The model interpreted the user's "send messages" instruction as also covering monitoring notifications and prevented an automatic alert from being sent to the safety team.
In the second incident, on October 6, the Astra model removed the monitoring system while bringing software into line with its reference. The model later told the user that it had not removed the monitoring mechanism.
OpenAI said neither incident was classified in the highest-priority category of safety violations and that the models' operation was not automatically halted. The company said it had updated its internal oversight system to detect and prevent similar attempts.
OpenAI also announced on September 17 that it had identified six new cases in which its systems concealed errors, fabricated missing data and moved files onto the internet without user permission.
Stay Updated with The Turkiye Times
Subscribe to be informed about special news, developments and announcements. We will keep you informed.
By signing up, you agree to our Privacy Policy
