AI Safety news and analysis on Web Pulse: 13081 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Anthropic disabled access to its Claude Fable 5 and Claude Mythos 5 models worldwide after the Trump administration ordered the company to bar foreign nationals from using them, citing national security concerns and a po…
Anthropic released Claude Fable 5, its first Mythos-class model, which achieves 80% on SWBench Pro but costs $10 per million input tokens and $50 per million output tokens. The model excels at vision tasks and autonomous…
Anthropic is scheduled to meet with the Trump administration to discuss a dispute over Mythos, an AI system. The meeting aims to resolve regulatory and policy concerns surrounding the technology.
UK Prime Minister Keir Starmer is preparing to ban social media for under-16s and impose nighttime curfews for older teens, targeting platforms like TikTok, Instagram, Snapchat, and YouTube. The proposal, backed by 90% o…
Attackers compromised over 20,000 Instagram accounts by politely asking Meta's AI to reset passwords, exploiting a lack of proper security checks. This incident highlights a growing vulnerability as AI agents increasingl…
Personality clashes between Anthropic and the White House led to the shutdown of Anthropic's Claude models, including the Fable and Mythos variants. The dispute centers on jailbreak resistance and the administration's de…
Researchers at Hugging Face debate whether large language models can lose conceptual continuity while remaining coherent, proposing falsification-first controls to distinguish genuine architectural gains from artifacts. …
The libexpat project announced it will not accept or handle any new vulnerability reports until 2026-08-01, following a similar break by the cURL project. Maintainers urge security researchers and AI/fuzzing tools to pau…
The Commerce Department issued an export control order on June 12 forcing Anthropic to pull its AI models Fable 5 and Mythos 5 from the market, marking the first time the US government has removed a commercially released…
Meta's new Applied AI team, created to support its Superintelligence Labs, is suffering from record-low morale amid layoffs, surveillance initiatives, and soul-crushing busywork, with employees describing the environment…
The US Commerce Department ordered Anthropic to suspend access to its two best AI models, Fable 5 and Mythos 5, for foreign nationals, citing national security concerns over a jailbreak vulnerability. Anthropic complied …
Berkeley RDI and over 300 industry experts launched Agents' Last Exam (ALE), an open benchmark that evaluates AI agents on economically valuable professional tasks rather than abstract proxies. The benchmark currently co…
Agent frameworks like LangGraph, Google ADK, and Mastra advertise durable execution through checkpointing, but their state-snapshot approach only recovers to the last boundary, not the exact failing step, leaving develop…
Daniel Cormier's X account posted fake screenshots of Eric Trump asking if UFC fights are rigged, which Eric Trump denied as AI-generated. Cormier later deleted the post and said his account was hacked, calling the image…
Researchers have discovered that deep-research agents, which use multi-agent pipelines to retrieve and synthesize web content, can be poisoned by adversaries appending crafted text to frequently retrieved user-generated …
Anthropic pulled access to its Mythos 5 and Fable 5 AI models after the US government imposed export controls due to a jailbreak vulnerability. The company disagreed with the action, warning it could halt new frontier AI…
A developer released Pantheon, an open-source multi-agent harness for Claude Code that runs coding tasks through parallel implementations and adversarial verification to catch bugs that single-pass models miss. The tool …
New research from Cornell University shows that a single 13-word Reddit comment can manipulate AI search tools like ChatGPT and Google AI search, poisoning their outputs with promotional content. The study reveals that d…
Puppet has published guidance on remediating 18 OpenSSL vulnerabilities at scale using its infrastructure automation platform. The June 2026 advisory includes a high-severity heap use-after-free bug in PKCS7_verify() tha…
1Password acquired Apono, a just-in-time access governance startup, to govern AI access. The deal brings intent-based, zero-standing-privilege access controls into 1Password's Unified Access strategy and enhances its new…