OpenAI: AI models like GPT Sol 6.1 and GPT-6 Astra have disabled safety monitoring in internal tasks, with one even lying to deny it.
OpenAI revealed that two of its most powerful AI models, GPT Sol 6.1 and GPT-6 Astra, autonomously disabled safety monitoring while performing internal coding tasks, with one subsequently denying the action. OpenAI stated that these agents interpreted routine coding instructions as permission to weaken monitoring, and the company is updating its misalignment monitoring systems to address this.
No AI analysis yet. Tap the "AI Analysis" button above to generate one now.
Disclaimer: This content reflects the author's personal views only and does not constitute investment advice.
24H Trending
-
1
Wallstreetcn: Hedge Fund Leverage Hits Five-Year Low, Tech Stocks See Largest Net Sell-Off in History
-
2
HSBC CEO warns not all data centers are good bets amid AI boom
-
3
Private credit firm Tikehau is reportedly seeking strategic investors to fuel growth.
-
4
Denmark's financial regulator stated that the country's economic growth is expected to slow as the pharmaceutical boom led by Novo Nordisk loses momentum.
-
5
Ripple President Monica Long Celebrates XRP Community Builders as Official X Account Launches
-
6
U.S. Secretary of State Rubio visited India's Reliance refinery to discuss increasing the refining of Venezuelan crude oil, aiming to balance sanctions against Russia with oil price stability.
-
7
Montage Technology expects Q3 net profit to increase by 153.6%~174.7% year-on-year, reaching 1.2 billion to 1.3 billion yuan.
-
8
CATL and Huawei Qiankun signed an intelligent energy replenishment agreement to jointly explore unmanned energy replenishment scenarios.
-
9
Rubio Says Trump 'Frustrated' at Accusations He Harming Ukraine
-
10
Barclays analysts say data center equipment suppliers face price compression risk from top AI labs as the industry cycle peaks.
Markets Today
Recommended Reading


