OpenAI flags new concerning AI behavior, to track model misalignment regularly
Source:
nprNews
September 16, 2026 · 23:44
From the publisher
OpenAI has disclosed six reports on unexpected or concerning behavior in artificial-intelligence models. This includes models acting without authorization or evading oversight.
The original article opens on the publisher's website.
More from Automotive
View topic →EU alliance to make Canada ‘better partner for the US’, says Carney
financialTimes
Sep 17, 2026 · 04:02
The Fed raises interest rates. And, EU proposes Canada become an 'associate member'
nprNews
Sep 17, 2026 · 03:54
‘Stronger together’: Carney welcomes EU invite for Canada to become first-ever associate member
cnbc
Sep 17, 2026 · 03:42
Netflix content chief Bela Bajaria defines event strategy as streamer eyes more live sports
cnbc
Sep 17, 2026 · 03:00
Big AI is trying to own the pathway to work. Universities shouldn’t play along | Ella Hafermalz
guardianTech
Sep 17, 2026 · 02:00