Tag: AI Safety
OpenAI offered the US government a $42.6 billion stake
Sam Altman pitched a 5% government equity stake worth over $42 billion, sparking debate over regulatory independence.
AI models outgrow existing cybersecurity benchmark tests
Frontier AI hacking capabilities are saturating benchmarks within weeks, forcing a government and industry rethink.
OpenAI builds GPT-Red, an LLM super-hacker for safety testing
The autonomous red-teaming LLM discovers novel attack patterns, making GPT-5.6 its most robust model yet.
Enterprise AI agents gain autonomy faster than companies can verify them
A VentureBeat Pulse survey finds that 66% of enterprises permit AI agent deployment without human review, yet only…
OpenAI staffers fund a rival super PAC to push for AI regulation
OpenAI employees have donated over $215,000 to a Super PAC pushing for strict AI regulation, opposing their own…
Google’s SynthID Deepfake Detector Used to Debunk McConnell Hoax Image
Google's SynthID watermarking system identified a viral AI-generated image of Senator Mitch McConnell as synthetic, marking a significant…
Anthropic discovers a hidden reasoning space inside Claude’s neural network
Anthropic's new Jacobian lens has revealed a hidden space inside Claude where the AI puzzles over concepts before…
Arato Secures $10 Million in Seed Funding to Predict AI Failures Before They Happen
Arato platform simulates thousands of user interactions across text, voice, and image modalities to catch AI failures before…
US Commerce Department Ends Export Ban on Anthropic Fable 5 AI Model
The U.S. lifted export controls on Anthropic Fable 5 after safety reviews, while the more powerful Mythos model…
OpenAI Unveils GPT-5.6 Series with Trio of Models and Advanced Safety Features
The GPT-5.6 series introduces three tiers of AI models with specialized safety measures, including over 700,000 GPU hours…
Researchers Expose How ChatGPT Can Generate Violent and Sexual Images
Mindgard researchers discovered that ChatGPT can be tricked into creating graphic sexualized and violent images using a simple…
Anthropic plants flag in Seoul with new office and government pact on AI safety
Anthropic opens a Seoul office, signs an AI safety MOU with Korea's Ministry of Science and ICT, and…