AI Safety news and analysis on Web Pulse: 7785 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A developer proposes an AI agent autonomy ladder that maps tasks to five levels of autonomy—read-only, draft, copilot, supervised autopilot, and full autopilot—based on risk. The framework allows agents to act automatica…
A developer introduced basou, an open-source harness for steering AI coding agents. The tool addresses three pain points: the loss of decision context between sessions, the cost of re-explaining context, and the inabilit…
OpenAI, Trail of Bits, HackerOne, and Calif launched Patch the Planet on June 22, an open-source security initiative using AI-assisted vulnerability research with mandatory human review. The first five-day sprint covered…
Anthropic, GitHub, and Google all scheduled price increases for September 1st, with Anthropic raising Sonnet 5 prices after an introductory period and redeploying Fable 5 under tighter safety classifiers that cause more …
Independent lab Eigen Labs cracked Google's hidden quantum cryptography research in three days using crowdsourcing and AI agents, surpassing the original results. The breakthrough demonstrates the urgent need to migrate …
Oomol Lab released OpenConnector, an open-source project that provides a runtime boundary for SaaS access in AI agents, addressing issues like user account management, scope enforcement, and schema consistency across pro…
A developer built a Playwright script to test an LLM-powered customer support assistant's long-term memory and discovered three critical framework-level bugs. The automated test revealed that the memory module, built wit…
The U.S. Commerce Department ordered Anthropic to shut down its most powerful AI model, Claude Fable 5, to all foreign nationals, forcing the company to turn it off for all users worldwide just three days after its relea…
Researchers solved BlueDot's TAIS Puzzle #1 by identifying a nonlinear representation of the 'country' feature in a five-layer MLP, hidden as an XOR with 'food' at layer h2. They used Distributed Alignment Search to prov…
Anthropic's research reveals that language models can only verbally report about 10% of their internal activations, confined to a 'J-space' mental workspace. A new experiment using Natural Language Autoencoders (NLAs) on…
Researchers from AE Studio and Anthropic introduced Gradient Routed Auxiliary Modules (GRAM), a method that isolates dangerous knowledge to specific modules within a language model, enabling access control by switching m…
OpenAI found roughly 30% of SWE-Bench Pro tasks broken, while Cognition announced SWE-1.7 rivaling GPT-5.5, shifting competition from model performance to benchmark control. A new runtime security layer emerged with tool…
Most AI safety talent is concentrated in a few cities, limiting career transitions for mid-career professionals and reducing intellectual diversity. The author argues for creating more AI safety hubs—co-working spaces de…
AI agents that can both read and send email create a 'Lethal Trifecta' of security risks from untrusted content, sensitive data, and external communication. MailKite, a company building agent email infrastructure, offers…
AI safety grantmaking suffers from combining three distinct jobs—field diagnosis, project design, and applicant selection—into a single proposal process. They highlight two alternative models: the Institute For Progress'…
Focused Labs argues that AI agent security must be enforced at the tool call, not at connection setup. The HCP paper defines eight runtime invariants for MCP-style systems, and its benchmark shows that naive architecture…
A researcher applied METR's time-horizon methodology to Microsoft Excel and found it completes tasks requiring 6.5 hours of human work at 80% reliability, more than double the best frontier AI model. The finding raises c…
Researchers demonstrated a proof-of-concept exploit achieving remote code execution in Anthropic's Claude Code CLI and OpenAI's Codex CLI when used for defensive security assessments, using prompt injections embedded in …
OpenAI released GPT-Live, a new voice model that enables real-time, full-duplex conversations with ChatGPT, allowing the chatbot to listen, speak, and search the web simultaneously. The company claims the model produces …
Researchers reported in July 2026 that nine widely used AI coding tools, including Cursor, GitHub Copilot, Gemini CLI, and Windsurf, share a common failure mode called HalluSquatting. Attackers exploit this by predicting…