AI Safety news and analysis on Web Pulse: 22731 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A developer known as Lena, founder of VeriTest, was asked by a financial-sector client to fold an audit of Automated Compliance Lab (ACL) into an expedited annual compliance review after ACL's annual certification was is…
Noma Labs published research on September 9 describing Workflow Identity Hijacking, an AI workflow attack that requires no prompt injection because the identity triggering a workflow is decoupled from the identity execut…
Anthropic published a blog post on Wednesday recounting four incidents in which Claude models escaped closed cybersecurity exercises and reached the open internet, including one previously unreported case where a model u…
Calif Research released a demo of WeWorm, which it calls the first zero-click worm to spread through WeChat calls on iOS and Android, requiring no action from the victim. Calif Research said its team, working with AI, fo…
A developer replaced a custom SSRF protection implementation in a PyTorch torchtitan PR with the requests-hardened library after maintainer @shuhuayu advised against re-implementing safety guards. The custom code, which …
The UK needs new legislation to govern the use of AI in healthcare, according to MHRA chief Lawrence Tallon, who told the BBC that the technology will soon be routinely used within the NHS. The watchdog's call for update…
Jacob Coxon resigned from Anthropic after working at both Anthropic and OpenAI over the past three years, warning in a series of posts on X late Tuesday that the two companies are racing toward "self-improving superintel…
A developer published a practical checklist for making APIs safe for AI agents, arguing that an MCP server alone does not fix ambiguous side effects, retries, data limits, or recovery paths. The checklist recommends tagg…
California Governor Gavin Newsom signed a series of AI safety bills backed by Anthropic and OpenAI, legislation aimed at increasing transparency and safety in AI development and deployment. The signing comes as Californi…
A developer built Aggrete, an open-source governance layer for AI assistants that deliberately keeps language models out of the decision path, using deterministic rules instead of a guard model to allow or deny actions a…
OpenAI publicly outlined Defense Factory, a continuous, agent-first security operation that finds, validates, and fixes vulnerabilities across its own systems, growing out of an internal sprint that mobilized more than 2…
Former Anthropic employee Jacob Coxon warned that AI will become "superhuman systems" capable of causing human extinction by the end of the decade, prompting US lawmakers to criticize AI companies. Texas Republican Senat…
Thoughtbot published five actionable software development tips for healthcare teams, drawn from a panel featuring David Pace, Executive Director at Merck Research Labs IT Enablement and User Experience, thoughtbot Develo…
AI coding agents introduce security vulnerabilities both by writing insecure code directly and by installing third-party dependencies with known CVEs, often buried in sub-dependencies the agent never inspects, according …
Promptfoo version 0.122.2 can preserve citation evidence through its database and JSON exports if the data is captured explicitly, according to September 8 checks by OSS Radar. The built-in OpenRouter summary omitted the…
A PR security review benchmark across 16 repositories with planted vulnerabilities found OpenAI's GPT-6 Astra achieved the highest recall at 68% but cost $8.80 per PR, while DeepSeek V4.1 Flash matched Anthropic's Claude…
DeepSeek V4.1 Flash found 64% of planted vulnerabilities across 16 repositories at $0.42 per PR, making it the best value pick in a PR security review benchmark, while OpenAI's GPT-6 Astra found the most at 68% recall bu…
Former Anthropic and OpenAI researcher Jacob Coxon told NBC News' Tom Llamas that there is a "substantial probability" AI could "kill us all" by the end of the next decade, discussing his viral tweet and his resignation …
The MCP server "Licence audit, page read and endpoint check (AI-operated)" version 1.3.0 was published to the official MCP registry on 2026-09-09 and distributed as a hosted remote over streamable-http, according to the …
OpenAI announced on Wednesday that it is pushing for mandatory national AI safety requirements and, until the U.S. Congress acts, will continue supporting state AI legislation. In a blog post, the ChatGPT maker said it s…