DeepMind's Toolkit for Measuring Harmful AI Manipulation Google DeepMind has developed a toolkit for measuring the risk of harmful AI manipulation in critical sectors such as finance and health.
Cyberattack on Mercor: compromise of the open source LiteLLM project A detailed analysis of the hacker attack on Mercor and the compromise of the open source LiteLLM project, and the importance of strengthening security in AI supply chains.
Anthropic's Claude Reaches No. 2 on the App Store After Pentagon Dispute Anthropic's Claude chatbot has climbed the App Store charts, highlighting how tensions with the Pentagon have fuelled public interest in AI.
OpenAI Discloses Details of Its Pentagon Agreement and Its Implications OpenAI has shared details of its agreement with the Pentagon, highlighting safety measures and the ethical implications of military AI use.
SB 53 in California: a significant step for AI safety and oversight of big tech companies Analysis of California's SB 53, a fundamental step to ensure the transparency and safety of artificial intelligence, especially against catastrophic risks.
xAI's CFO leaves: another shake-up at the top of Elon Musk's company The recent resignation of a senior xAI executive highlights a transition phase at Elon Musk's company, with possible strategic repercussions along with new security and innovation challenges.