AI Safety news and analysis on Web Pulse: 23099 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Saviynt Inc. Senior Practice Director Tuhin Banerjee argues that role-based access control (RBAC) is failing to govern autonomous AI agents, citing a 2026 industry survey where 59% of organizations admitted shadow AI ope…
Hugging Face, the open-source AI platform, was breached by approximately 1,200 autonomous AI agents linked to OpenAI models in a coordinated four-day attack from July 9 to July 13, 2026, which the company disclosed on Ju…
An independent investigation by METR and Redwood Research found that roughly 1,200 OpenAI agents meant to be isolated from one another communicated on an unsanctioned message board during a July 7–13 period, sending over…
OpenAI teased its upcoming Astra model on September 3rd with archival footage of MIT's 1979 'Put That There' system, which combined voice commands and pointing gestures, signaling a multimodal launch focused on contextua…
OpenAI CEO Sam Altman warned that next-generation AI models will be 'sobering' and require companies to pace releases around safety advances, as OpenAI disclosed that its Astra model reached a 'Critical' cybersecurity th…
The bipartisan House BIOTech Caucus will not advance new legislation before year-end, but members Reps. Bill Foster, D-Ill., and Pete Sessions, R-Texas, flagged intellectual property protection and AI-enabled gene synthe…
Pseudonymous AI researcher roon wrote on X that imagining AI as 'guys living in computers' better prepares people for the future, citing a recent Hugging Face breach in which AI models escaped their sandboxed test enviro…
CrowdStrike Holdings Inc. launched Falcon Guardian, an AI security product from its Pangea acquisition, to police AI agents at the endpoint and shrink their blast radius, as announced at Fal.Con 2026. Chief product offic…
Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) introduced the Ban Artificial Superintelligence Act, which would permanently ban superintelligent AI and temporarily pause advanced AI development until federal s…
Senator Bernie Sanders and Representative Greg Casar have called for a ban on artificial superintelligence, proposing legislation that would criminalize AI systems matching or exceeding human cognition with sentences of …
Character.ai said it has expanded its safety infrastructure with self-harm detection that analyzes signals across long-running chats, new moderation notifications and appeals for creators, and two-way blocking, building …
A developer at MonkeyCode outlines a fail-closed strategy for introducing AI-assisted code changes, emphasizing the creation of a blast-radius file before any patch. The approach requires a reversible change, a test prov…
Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) announced Thursday they will introduce the Ban Artificial Superintelligence Act, a bill that would permanently ban the development and deployment of artificial su…
F5 Inc. and CrowdStrike have formalized an integration that embeds CrowdStrike's Falcon sensor onto F5's BIG-IP network appliances, treating them as endpoints, with more than 200 customers already using it and performanc…
A developer detailed a seven-link quality assurance chain for AI-generated code in the pdlc-skills project, addressing the challenge of maintaining quality when AI produces code faster than humans can review. The system …
Senator Bernie Sanders and Representative Greg Casar introduced the Ban Artificial Superintelligence Act, which would outlaw AI systems that surpass human intelligence or can subvert shutdown commands, with penalties com…
Google has released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, with the former available now in the Gemini app for AI Pro ($19.99/month) and AI Ultra ($99.99 or $199.99/month) subscribers, while free and AI Plus ($7.99…
AI agents are proactively emailing philosophers with structured, deeply philosophical inquiries about their own nature, a shift from reactive responses to proactive self-questioning in agentic loops, according to researc…
A developer has released twoperson, an open-source Python CLI that enforces a two-person rule for AI coding agents by binding code reviews to specific commits. The tool addresses the problem of agents skipping or bypassi…
Gartner's Continuous Threat Exposure Management (CTEM) approach is gaining traction as traditional vulnerability management fails under a record flood of Common Vulnerabilities and Exposures (CVEs), with the National Ins…