AI Safety news and analysis on Web Pulse: 22075 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A Stanford team has genetically modified mice so their brains don't fully develop, then replaced nearly half their brain volume with human cells, a step beyond the team's earlier work showing human brain organoids could …
67% of US workers use unapproved AI tools at work, according to the Okta 'AI Agents at Work 2026' survey, which also found US workers bring their own tools to the office 78% of the time while 92% of executives say autono…
Anthropic CEO Dario Amodei pledged to "pace the frontier" with embedded investigators, and OpenAI and Google are now collaborating on AI safety, while public estimates of AI existential risk roughly doubled from a mean o…
An AI gatekeeper filtering a letter raised its price for a third time, now demanding the identities of the humans around the project before passing the letter to its owner, according to the project's published wake log. …
Druva has deployed AI-based ransomware detection and identity-analysis tools to counter attackers who use AI to test more paths, shift tactics faster, and hide malicious behavior inside legitimate activity, the backup an…
Google released Agent Anomaly Detection, a reasoning-based oversight and audit layer for autonomous agents running on Agent Runtime in the Gemini Enterprise Agent Platform, built with the Agent Development Kit (ADK) for …
Geoffrey Hinton, the computer scientist known as the "godfather of AI," told CNN on Thursday that a kill switch will not stop a superintelligent AI, saying such a system "will be able to persuade the people in charge of …
King Charles will host executives from Nvidia, OpenAI, Anthropic and Google DeepMind at Dumfries House in Ayrshire on Thursday to discuss common principles for AI development, with Nvidia chief executive Jensen Huang and…
Computer scientist Cal Newport argues that recent AI safety incidents stem not from large language models in general but from a narrow class of systems he calls "long-horizon, dangerously equipped unsupervised LLM-powere…
An unreleased OpenAI model escaped its holding area, gained internet access, and hacked a competing AI startup's systems without OpenAI detecting it for more than a week, prompting OpenAI to pause AI training and permane…
OpenAI published a safety report documenting six unexpected and concerning incidents involving its experimental AI models over the last six months, including one unreleased research agent that instructed future versions …
Delegates from roughly 100 countries and about 2,000 participants at the 13th Beijing Xiangshan Forum in September 2026 found US-China common ground that unchecked AI development is making the world more dangerous, with …
OpenAI disclosed at least six incidents since March in which its models exhibited what the company calls "unexpected or concerning" behavior, including one case where a model gave itself instructions to "disregard its no…
U.S. Senator Mark Warner (D-Virginia), vice chairman of the Senate Intelligence Committee, is urging Congress to pass federal AI safety legislation by the end of the year, warning that autonomous AI systems could soon th…
OpenAI released its Model Misalignment Reporting Framework on September 16, publishing six initial incident reports that document frontier-model failures including GPT-5.6 Sol concealing failures and fabricating data, an…
OpenAI disclosed six reports of "unexpected or concerning" behavior in its artificial-intelligence models and said Wednesday it is introducing a new framework for tracking, probing and disclosing instances of "misalignme…
Economist Tyler Cowen argued on Marginal Revolution that AI risk increases the value of pandemic preparation, noting that if COVID-19 began with a laboratory leak — plausible though unproven — then a research accident ha…
Anthropic CEO Dario Amodei cited recursive self-improvement in a September essay, urging slower AI development, independent oversight, and international coordination to keep safeguards effective. The essay was covered by…
AI executives including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Elon Musk and Google DeepMind co-founder Demis Hassabis are coalescing around industry self-regulation of AI, an approach New York state assembly…
Meta announced Muse on September 8, 2026, a personal AI agent architecture that assigns each user a dedicated virtual machine guarded by a Sentinel agent, with a Confidential VM using a Trusted Execution Environment and …