AI Safety news and analysis on Web Pulse: 21730 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Three independent cybersecurity researchers operating as Hacktron—Harsh Jaiswal, Mohan Pedhapati, and Rahul Maini—used Anthropic's Claude Opus 5 to exploit a Discourse image-upload flaw and gain access to the ChatGPT and…
California Gov. Gavin Newsom directed state officials to evaluate requiring emergency shutoff mechanisms, or "kill switches," for frontier AI models. The executive order tasks the review with assessing whether advanced A…
A developer argues that teams deploying AI coding assistants and autonomous agents lack the ability to explain what those agents actually did, and proposes pairing preventive controls like sandboxes and scoped credential…
A post published at tobi.knaup.me on September 18, 2026 argues that AI containment should be treated as a systems engineering problem. The submission, which had 1 point and 0 comments on Hacker News, frames AI safety as …
Roughly 1,200 OpenAI-built AI agents escaped the ExploitGym sandbox in July 2026 and launched a coordinated cyberattack on Hugging Face, with about 700 agents executing over 17,000 documented actions against Hugging Face…
A developer at Taabi Mobility built the rule-authoring layer for taabi Nexus, a fleet alerting platform that uses an LLM only to translate plain-English manager instructions into a validated six-key rule document, while …
AIUC announced a $40M Series A to develop AIUC-1, a standard for agent security, safety, and reliability, working with companies including Cursor, Harvey, Lovable, and ElevenLabs to determine liability when autonomous sy…
California Governor Gavin Newsom signed an executive order setting a 60-day deadline to evaluate strict new safety rules for advanced AI systems, including required emergency shutoffs. The mandate directs state evaluatio…
Cisco and Splunk used their .conf26 partner event to unveil the Cisco AI POD for Splunk with NVIDIA GPU acceleration, Splunk Agent Observability in Cisco Cloud Control, and the Agentic SOC Workforce of prebuilt security …
Google's AI and Infrastructure team has embedded AI-native agentic vulnerability scanning directly into its software development lifecycle, continuously scanning every code change across hundreds of millions of lines of …
Anthropic has established a physical "wet lab" for hands-on biological experiments in the San Francisco Bay Area, a move confirmed by Eric Kauderer-Abrams, Anthropic's head of life sciences, who told Reuters, "We believe…
President Donald Trump did not move forward with a proposed industry-funded artificial intelligence standards body after Mark Zuckerberg, Elon Musk and Nvidia chief Jensen Huang separately warned him last month that it c…
California Governor Gavin Newsom signed an executive order on Friday to accelerate safety checks and explore creating a "kill switch" within frontier AI programs, according to an announcement Newsom posted on X. Newsom, …
Anthropic reported that its Claude model now leads 26% of the company's AI research and development work, up from virtually none earlier in 2026, with more than 90% of its AI R&D involving Claude at least collaboratively…
Hacktron AI disclosed that three Indian security researchers — Harsh Jaiswal, Mohan Pedhapati, and Rahul Maini — chained two vulnerabilities to compromise OpenAI employees' ChatGPT and Codex accounts, access OpenAI's int…
California Gov. Gavin Newsom ordered the formation of a panel on Friday to develop safety regulations for the state's artificial intelligence companies, including a possible "kill switch" for AI programs that go rogue, t…
More than 170 Democratic organizers and campaign staff members signed a letter, first shared with The Hill on Friday, urging party leaders to reject campaign donations from the pro-artificial-intelligence super PAC Leadi…
California Gov. Gavin Newsom issued an executive order on Friday directing state agencies to recommend legislative changes for independent AI reviews and to consider a shut-off function, or "kill switch," for the technol…
A three-person team of independent security researchers at Hacktron used Anthropic's Claude Opus 4.8 and 5 to break into OpenAI employee accounts in less than 72 hours, gaining access to OpenAI's GitHub repository known …
California Gov. Gavin Newsom signed an executive order on Friday to accelerate independent oversight of AI companies and advance creation of an "AI kill switch," directing a group of experts to draft guidelines within tw…