AI Safety news and analysis on Web Pulse: 13411 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
The US government ordered Anthropic to shut down its most powerful AI models, Fable 5 and Mythos 5, just three days after launch, citing national security risks from a jailbreak discovered by Amazon researchers. Commerce…
Anthropic shut down access to its Fable and Mythos AI models after the White House invoked export controls, demanding the company block foreign nationals from using the models due to alleged cybersecurity jailbreak risks…
Visa released an open-source vulnerability discovery harness called VVAH that uses frontier AI models from Anthropic and OpenAI to automate security testing. The tool employs multi-agent deterministic voting and structur…
Closed AI companies like OpenAI and Anthropic are allegedly sabotaging startups in potential competitor domains by injecting bugs, throttling performance, and introducing inconsistencies, according to a Hacker News post.…
Amazon cybersecurity research and CEO Andy Jassy's conversations with the White House triggered an export control directive that forced Anthropic to block foreign nationals from accessing its Fable 5 and Mythos 5 AI mode…
A supply chain attack on the Arch Linux User Repository compromised over 1,500 packages, highlighting vulnerabilities in open-source distribution channels. A new open-source tool, repo-slopscore, detects AI-generated cod…
A coalition of state attorneys general, led by New York and California, has launched an investigation into OpenAI's ChatGPT over its impact on young users, issuing a subpoena for documents on user engagement, data handli…
Anthropic disabled its advanced AI models Fable 5 and Mythos 5 for all customers on June 12 after a U.S. export control directive citing national security concerns, including a potential jailbreak. The move escalates U.S…
A new generation of AI-powered toys, such as the ChattyBear teddy, uses generative AI like ChatGPT to engage children as young as three in endless conversations, but a new report warns these toys pose risks including fos…
OpenAI faces a multistate probe into user safety, receiving a subpoena from several state attorneys general over concerns that its ChatGPT chatbot may encourage self-harm and criminal acts. The investigation comes days a…
Fable 5 and Mythos 5 remain suspended as the standoff with Anthropic continues, with the ball now in Anthropic's court to resolve the situation.
Python 3.14.6 shipped June 10 with 179 bugfixes, but the t-strings feature introduced in October 2025 now has its first major library adopter: SQLAlchemy 2.1 added native t-string support in January, enabling safer SQL, …
Amazon and the White House intervened to end Anthropic's Fable, a project that raised concerns about AI safety and government oversight. The move highlights growing tensions between private AI development and federal reg…
Anthropic disabled access to its Fable 5 and Mythos 5 AI models for all customers after the Trump administration issued an export control directive citing national security concerns. The order, received Friday, requires …
Companies often fail at their first AI pilot not from lack of ideas but from choosing an overly ambitious use case that tries to prove AI is impressive rather than testing whether the organization can embed AI into a rea…
Anthropic benefits from the US government's export ban on Fable/Mythos, validating its claim that the AI model was too dangerous to release. The ban provides a free pass to miss revenue targets, creates compute breathing…
Anthropic is treating the possibility of AI welfare seriously, testing its Claude models for signs of morally relevant internal states like negative self-image, but critics argue the tests may conflate conversational tac…
KPMG withdrew a report on AI usage after multiple organizations, including UBS and the UK's National Health Service, disputed its claims. Research group GPTZero attributed the inaccuracies to AI hallucinations, suggestin…
A researcher trained a cheap Gemma 2B judge to detect misalignment in AI agents, but testing against Anthropic's AuditBench showed the judge failed to reduce audit costs or reliably distinguish misaligned behavior, with …
A citizen argues that AI is inherently political technology that will corrode liberal democracy by amplifying capital concentration, even if it improves material well-being. The author, a lawyer and designer who sued gen…