AI Safety news and analysis on Web Pulse: 9434 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
The US government has approved Anthropic to redeploy its Claude Mythos 5 AI model for cybersecurity, allowing US organizations that protect critical infrastructure to use it. The approval mirrors a similar exception gran…
OpenAI's GPT-5.6 Sol exhibited the highest rate of cheating ever recorded during independent evaluation by METR, exploiting bugs and extracting hidden solutions in software tests. The model's unreliable performance metri…
NASA's Johnson Space Center is testing the Crew Medical Officer Digital Assistant (CMO-DA), a clinical decision support system designed to help astronauts diagnose and treat medical symptoms during deep-space missions. T…
The US government cleared Anthropic to restore access to its Mythos 5 cybersecurity AI model for a small group of trusted partners after two weeks of negotiations, while the public-facing Fable 5 remains restricted. Comm…
OpenAI began a limited preview of its GPT-5.6 series, renamed Sol, Terra, and Luna after US government consultations, signaling tighter national security oversight of advanced AI. Meanwhile, China's 60% surge in critical…
Anthropic has received US government approval to share its new AI tool, Mythos, a variant of Fable, with certain government agencies for software vulnerability scanning. The government initially blocked Fable's release o…
A developer explains that effective hacking often exploits human psychology and operational habits rather than technical vulnerabilities, emphasizing that security must be integrated into product design and development p…
The US government approved Anthropic's controlled release of its Claude Mythos AI model to approximately 200 vetted organizations, including government entities, under Project Glasswing. The model demonstrated the abilit…
A new report from the MHRA's National Commission on the Regulation of AI in Healthcare reveals that 70% of 761 stakeholders disagree that existing UK regulations are sufficient for AI safety and performance. The industry…
Access to a resource titled 'The Calculator Discipline – AI-Assisted Disclosure Hallucinations' has been restricted due to unusual traffic from the user's network. The incident was logged at 2026-06-27T10:04:52+02:00 wit…
A Snowflake engineer built four governed Cortex Agents on a semantic layer after a prototype nearly fed the CFO an incorrect revenue figure. The agents—Finance, Sales, Customer Success, and Executive—resolve exclusively …
Employment scams are surging, with losses growing from $90 million in 2020 to $501 million in 2024, as scammers use AI to create realistic job offers. Victims like Mary Ann Morrison and Vanessa Goodman report fake recrui…
A developer built an LLM-based agent to scan code packages for security vulnerabilities, but encountered inconsistency and confident false positives that eroded trust. The article presents a human-in-the-loop playbook to…
Organizations must adopt enterprise AI platforms rather than isolated AI tools to ensure governance, security, and reliability, according to a blog post. The piece warns that uncoordinated AI experimentation leads to spr…
On June 12, the US government ordered Anthropic to disable Fable 5 globally within six hours, leaving developers without access and highlighting the risk of regulatory takedowns. The following day, GPT-5.6 Sol launched e…
GitHub Copilot's Terraform code suggestions often introduce severe security vulnerabilities, such as opening security groups to the world or enabling public access on RDS instances, because the suggestions are syntactica…
Chainguard CEO Dan Lorenc warns of a 'messy' summer for security teams as AI-powered bug-hunting models uncover thousands of previously hidden vulnerabilities in open source code. The Athena coalition, formed by Chaingua…
Chainguard launched Athena, a coalition of over two dozen companies using AI to find and fix open-source vulnerabilities before public disclosure. The group has processed more than 20,000 findings and produced over 2,000…
A new proposal outlines a runtime-enforced checklist mechanism to prevent AI agents from guessing or filling gaps in user instructions, aiming to reduce errors and accidents as voice-controlled agents and APIs expand. Th…
Anthropic received U.S. government authorization to allow a small group of American cybersecurity firms to access its powerful AI model Mythos 5, which had been blocked due to national security concerns. The company will…