OpenAI tightens controls on its new Astra model and halts some internal activities over cybersecurity risks, as it cannot rule out the model has reached "Critical" capability to launch auton
OpenAI表示,其初步评估显示该模型性能强劲,无法排除其已达到“关键”能力水平,即无需具体提示即可自主发起针对复杂网络防御的网络攻击。此举正值AI安全辩论日益激烈,此前Anthropic和Meta的AI系统也曾卷入安全事件,美国议员正推动引入“AI杀戮开关”法案。
Source:CNBC头条 · Source Link
Disclaimer: This content reflects only the author’s personal views and does not constitute any investment or financial advice. If you discover any content that violates regulations,Click to Report
24H Trending
-
1
Ethereum's Latest Trends and Future Directions: Modularity, Layer 2 Scaling, and Key Upgrades
-
2
How to Buy, Sell, and Trade KAVA Coin? A List of Platforms Supporting KAVA Trading
-
3
Kalshi prediction market valuation soars to $22 billion, institutional demand drives strong growth
-
4
Wall Street sees fund manager Scott Bessent的举动表明其渴望抑制债券收益率飙升,反映出对债市的担忧
-
5
Stock Funds Up 10.6% Year-to-Date in 2026 Despite Turbulent July
-
6
Cryptocurrency Perpetual Contracts and Leveraged Trading: Mechanisms, Risks, and Choices
-
7
Analysis of the Current Situation and Risks of USDT Transactions for Users in Mainland China
-
8
Over 100 Crypto Projects Fold in 2026 Amid "Dot-Com Style" Shakeout
-
9
Bank of Japan (BOJ) expects Japan's economy to grow moderately in FY2026, accelerating from FY2027 as crude oil price effects wane, supported by AI demand
-
10
LUNC Coin Latest Price Trends: An Analysis of the Terra Classic Community's Reconstruction and the Possibility of Returning to $1
Markets Today
Recommended Reading








