AI Safety news and analysis on Web Pulse: 12918 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A developer built a lightweight PII detection utility that runs inside InterSystems IRIS using Embedded Python, eliminating the need to export data for analysis. The tool scans tables, identifies PII via Microsoft Presid…
A developer found that LLM API providers can swap in cheaper models undetected, and the intuitive method of flagging low-quality responses fails because cheaper models produce more predictable text with lower perplexity.…
Ledgerline, a local-first finance MCP server, gives AI agents read-only access to financial data via a single SQLite file without cloud storage or exposing credentials. The tool supports bank sync through SimpleFIN Bridg…
Anthropic released Claude Fable 5 on June 9, but the U.S. Commerce Department issued an export-control directive barring foreign nationals from accessing it, forcing Anthropic to take both Fable 5 and Mythos 5 offline. T…
The G7 leaders' lunch with AI executives in Evian follows the US forcing Anthropic to disable its most advanced AI models, sparking global concern. In France, politicians and presidential candidates are calling for Europ…
Guillermo del Toro warned that AI and business consolidation are threatening creative freedom, stating 'We are on the verge of cinema illiteracy' during a BFI America dinner where he received the BFI Fellowship. The Osca…
The U.S. government's order for Anthropic to block non-U.S. access to its Fable and Mythos models is boosting adoption of open-source alternatives, especially from Chinese AI labs like DeepSeek and Knowledge Atlas. Chine…
AM Best warned U.S. property/casualty insurers that AI-driven data center proliferation creates underwriting risks beyond what the industry has previously faced, identifying business interruption as the most consequentia…
Code strings in the iOS 27 developer beta indicate Apple may add a break reminder for extended Siri AI conversations, reading: "You've been in this conversation for [n] hours - consider taking a break. Siri is not a pers…
The U.S. Department of Justice intervened in an NAACP lawsuit against Elon Musk's xAI, arguing that shutting down unpermitted gas turbines powering its Colossus 2 data center would threaten national security because the …
Several Pasadena police officers were placed on administrative leave after new evidence emerged in a shooting where an officer was shot during alleged 'horseplay' with loaded guns. Police Chief Gene Harris announced an a…
The shift from vibe coding to agentic engineering marks a new phase where judgment, not code generation, becomes the competitive advantage. Incidents like the Tea dating platform breach and an AI agent ignoring safeguard…
Derbyshire Police launched a criminal investigation into one of its officers for allegedly using AI to fabricate evidential material across multiple cases, potentially amounting to perverting the course of justice. The o…
A writer argues against smart home technology, preferring manual controls and expressing concerns about reliability, privacy, and the potential for AI failures, citing examples like robot waiters malfunctioning and the r…
Mozilla CTO Raffi Krikorian argues that the U.S. government's shutdown of Anthropic's Fable and Mythos AI models via export controls was reckless and counterproductive, as it fails to enhance safety while demonstrating u…
Lawyers for former Lush CEO Andrew Gerrie have increased their litigation budget to £6.1m, partly because they will only allow senior lawyers—who charge the highest fees—to train an AI model for document review, excludin…
Anthropic faces significant regulatory risk ahead of its planned IPO after the U.S. government blacklisted its flagship AI models twice, including a recent export control that forced the company to take its newest models…
Robinhood is laying off 10% of its full-time employees, about 290 people, in a restructuring move that notably avoids blaming AI, unlike many other tech companies. CEO Vlad Tenev's note to employees did not mention AI, i…
PiLogic signed a two-year Cooperative Research and Development Agreement with the U.S. Air Force Research Laboratory to field-test its satellite fault-detection and failure-prediction software. The agreement focuses on s…
Okta expanded its partnership with Google Cloud to integrate identity governance into AI agents and enhance Chrome browser security. The initial release connects Auth0 for AI Agents to the Gemini Enterprise Agent Platfor…