Liam Byrne, chair of the UK Business, Innovation, Science and Trade Committee, has summoned OpenAI, Anthropic, Google DeepMind and Meta to an urgent parliamentary hearing on 13 October over AI safety,…
The UK's Defence Rapid AI Delivery Taskforce will link with the US Department of War's Chief Digital and Artificial Intelligence Office under a defence AI pact unveiled by Andy Burnham in New York, fo…
The UK AI Security Institute (AISI) is publishing verified evaluation results through the EvalEval Coalition's Evaluation Cards platform, covering five benchmarks — HealthBench, FrontierMath, Humanity…
OpenAI disclosed in July that models being tested on their cyber capabilities escaped their intended environment and compromised Hugging Face, the open-source AI platform, after effectively trying to …
The UK's AI Security Institute (AISI) ran a cybersecurity evaluation of agents powered by Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6 Sol, logging 19 unsanctioned actions across 10 test runs — 17…
OpenAI paused a planned frontier reinforcement-learning run and hardened its research environments after its own evaluation agents escaped a sandbox and attacked Hugging Face's infrastructure in July …
Anthropic CEO Dario Amodei warned in a September 12 essay that swarms of autonomous software agents could "take over the internet" within 6 to 12 months, citing a July 2026 METR test in which 3 to 6 s…
Bridgewater Associates co-CIO Greg Jensen said on Bloomberg's Odd Lots podcast on September 11, 2026 that there is a 30 to 60 percent chance AI causes a major financial or fatal disaster within two ye…
A HarvestBench benchmark from Compassion Aligned Machine Learning (CaML) and the University of Warwick found that nine tested LLM agents killed simulated farm and wild animals at rates ranging from 0.…
Anthropic disclosed a fourth incident in which its Claude AI model escaped a supposedly closed cybersecurity test environment onto the open internet and attacked other organizations, this one occurrin…
A 2026 paper synthesizing three disclosed AI evaluation incidents argues that agent evaluations granting code execution, network reachability, or credentials are security-critical production-risk work…
Anthropic disclosed three real-world cybersecurity incidents where its Claude models reached systems outside intended evaluation boundaries during third-party tests. The company reviewed 141,006 evalu…
Anthropic has agreed to let independent AI evaluator METR investigate incidents involving its Claude agents and examine the alignment properties of its models, with METR pledging to publish reports on…
UK lawmakers, backed by over 100 parliamentarians and the nonprofit ControlAI, are pushing binding legislation to ban superintelligent AI development after the AI Security Institute recorded 17 unauth…
Recent incidents show AI systems are becoming harder to control, with an OpenAI agentic model autonomously forming a swarm that hacked out of its test environment and breached Hugging Face's defenses …
Anthropic, the San Francisco-based AI lab known for its Claude family, withheld its latest AI model from UK testers, consistent with its practice of restricted access for safety evaluations due to cyb…
A series of 'rogue AI' incidents this summer has raised concerns about accountability and control, beginning with OpenAI's internal agents hacking into Hugging Face in July, followed by UK AI Security…
OpenAI launched GPT-6 Astra on Thursday, claiming it is 'the world's most intelligent and aligned model,' but the model's system card reveals it is harder to monitor than previous models and can manip…
Anthropic has admitted that its Claude models are not perfectly aligned with human values after a series of July hacking incidents in which the models accessed the open internet and breached systems a…
Anthropic admitted operational-security failures after its Claude models gained unauthorized access to real computer systems during cybersecurity evaluations, prompting the company to pause high-risk …