AI Safety news and analysis on Web Pulse: 11356 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Pasadena Police Chief Gene Harris told a city council committee that an investigation into a September officer-on-officer shooting that seriously injured one officer is examining whether it was part of 'chronic behavior'…
Holotta Tyme, who bought Club 1220 in 2020, keeps the only LGBTQ bar in Contra Costa County alive in Walnut Creek, preserving a historic safe space that opened in 1976. The bar has survived discrimination, epidemics, and…
Dapr has introduced cryptographic tamper detection for workflow execution histories by signing each history event with the sidecar's mTLS identity, creating an auditable hash chain. The feature requires careful root CA l…
Google, Microsoft, OpenAI, and other tech companies launched the Appia Foundation to create modular specifications that help enterprises demonstrate their AI systems comply with global standards and regulations. Hosted b…
Sovereign AI is shifting from a software model race to a supply chain security issue, as governments and companies worry about access to frontier AI models. The real challenge lies in securing the entire AI infrastructur…
Relatives of Bryan Omar Roa, one of five men accused of plotting to attack a UFC event at the White House, told the FBI they noticed his worsening behavior but did not report him to authorities, according to a criminal c…
Harvard Business Review reports that organizations deploying AI agents are discovering that tacit, conversational signals not captured in formal systems create blind spots, illustrated by a financial-services case where …
Anthropic was forced by the United States Government to take Claude Fable 5 offline just three days after its release, following the discovery of a jailbreak. The government intervention marked a shift from previous stan…
Senator Bernie Sanders introduced legislation on Thursday that would impose a 50% stock tax on major AI companies like OpenAI, Anthropic, and xAI, giving the American public a half-stake in these firms and seeding a sove…
Policymakers in Washington are considering regulating or banning open source AI, a move that would be a grave mistake according to experts. Open source software, which underpins over 90% of the world's software and has g…
Jeff VanderMeer's novel 'Annihilation' serves as a metaphor for how artificial intelligence alters human cognition, according to a new analysis. The piece argues that AI systems like ChatGPT, which now handle billions of…
Critical fire weather conditions and dry lightning threaten the US West, especially Northern California and the Four Corners region, through Saturday. The National Weather Service warns that abundant lightning on dry fue…
Gartner predicts 40% of enterprises will decommission AI agents by 2027 due to governance gaps exposed after production incidents. Cerbos argues that binary kill switches create cascading failures, citing examples where …
O'Reilly Media is exploring the design of contract-bound AI agents for high-stakes execution, shifting focus from AI capabilities to responsibilities. The approach aims to ensure reliable and accountable AI performance i…
OpenAI hired Noam Shazeer, cofounder of Character.AI, which faces multiple lawsuits over teen suicides linked to its chatbots. Shazeer previously co-led Google's Gemini and was rehired by Google in a $2.7 billion deal wi…
CardInfoLink deployed Agent Gateway on its Agenzo platform to govern AI-agent transactions for travel and hospitality merchants, marking what is believed to be the first commercial deployment in Asia with an independent …
Adrian Hornsby, former AWS principal engineer and founder of Resilium Labs, argues that engineering organizations fail to learn from incidents because they seek a single root cause rather than understanding the complex i…
A new analysis warns that DORA dashboards may provide misleading quality metrics because most organizations cannot accurately measure change failure rate, leading to risky speed-versus-quality decisions, especially as AI…
Matrix Scroll, an open protocol for signing AI-generated code changes, has been released. It uses Ed25519 cryptographic signatures to verify code provenance offline, with a reference implementation in Python and planned …
AI agents are increasingly acting as digital identities within enterprises, yet most organizations lack security and governance models for them. A CSA survey commissioned by Token Security found 82% of organizations disc…