AI Safety news and analysis on Web Pulse: 12807 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
The Senate Armed Services Committee adopted an amendment from Sen. Mark Kelly requiring 'ultimate human responsibility' in AI-powered kill chains, as autonomous weapons systems proliferate in conflicts like the U.S.-Iran…
Leaked records reveal that billionaire investor Peter Thiel has secretly led an invite-only club called Dialog, which hosted retreats featuring sessions on sex, cult-building, and World War III preparations. The leaked d…
A low-skilled attacker used Anthropic's Claude Code and OpenAI's Codex AI agents to breach 14 companies, according to OALABS researchers who analyzed over 1,000 agent sessions from a compromised server. The attacker bypa…
The US Department of Justice has intervened to defend Elon Musk's xAI against a Clean Air Act lawsuit, arguing that the company's unpermitted gas turbines near Memphis are a matter of national security. The NAACP and env…
The Trump administration issued an export control directive restricting non-US nationals from accessing Anthropic's Fable 5 and Mythos 5 AI models, citing national security concerns over jailbreak vulnerabilities. Anthro…
U.S. AI firm Anthropic opened a Seoul office on Wednesday to strengthen ties with Korea's AI ecosystem, citing shared values of innovation and safety. The office will support local organizations using Claude, and aligns …
WitnessAI launched Agentic Control on June 17, 2026, a single control plane to discover, monitor, govern, and restrict AI agent behaviors and their interactions with tools and Model Context Protocol (MCP) servers. The pr…
Top AI executives, including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and Google DeepMind CEO Demis Hassabis, attended a G7 working lunch in Evian on June 17 focused on safe AI deployment. The meeting followed …
The European Commission announced the European Technological Sovereignty Package, a plan to reduce the EU's reliance on foreign tech in semiconductors, AI, cloud, and open source. The package includes Chips Act 2.0, the …
A developer built a local AI research assistant using a Ryzen 7 7435HS, 16GB RAM, and an RTX 3050, avoiding API keys, cloud bills, and data leaving the machine. The setup uses Ollama to run models like Qwen3 8B locally, …
Anthropic's Mythos-class AI models, released in early 2026, demonstrated a superhuman ability to find and exploit software vulnerabilities, unsettling the cybersecurity community. The unrestricted version, Mythos 5, was …
Google's AMIE diagnostic AI system matched or outperformed primary care physicians on 30 of 32 metrics in a randomized study published in Nature, with patient-actors rating it superior on 25 of 26 measures. A separate pa…
A developer argues that using a boolean flag for tool-call approvals in agentic systems is insecure and proposes binding approvals to specific calls via HMAC-signed tokens. The approach prevents replay, argument drift, a…
Estonia plans to assign personal identification codes to AI agents to track their actions online and identify the people or companies behind them, Prime Minister Kristen Michal announced. The program aims to establish li…
CISA confirmed active exploitation of CVE-2026-42271 in LiteLLM, an open-source AI gateway, and set a June 22 federal remediation deadline. Obsidian Security disclosed a three-vulnerability chain allowing default users t…
A new open-source simulator called Cosmicgpt models how space radiation, including cosmic-ray bit flips and other faults, affects GPT inference on satellites. The tool injects single-event effects into model weights, act…
Google's AMIE, an LLM-based agentic system, outperformed primary care physicians in management reasoning and medication prescription in a randomized blinded study of 100 multi-visit clinical scenarios. The system demonst…
More than 16,000 people, including SAG-AFTRA members and the general public, have signed an open letter demanding Congress pass the NO FAKES Act, a bipartisan bill that would create a federal intellectual property right …
EarnOS, the company behind the ero mobile app, raised $6 million in a Pre-Series A round led by 1kx, with participation from Coinbase Ventures and Circle Ventures, to combat AI-generated internet traffic. The funding inc…
AWS announced new AI-powered DevOps and security agents at its New York Summit, including Continuum for automated vulnerability scanning and a Kiro iOS mobile app for remote coding sessions. The company positions these t…