AI Safety news and analysis on Web Pulse: 11241 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Norway will impose a near-total ban on AI tools in primary schools starting August 2026, with limited supervised use for older students, as part of efforts to combat negative impacts on learning. Prime Minister Jonas Gah…
An AI forecaster predicts the U.S. government will force Anthropic to restrict Claude Fable to non-Americans, setting a major precedent for AI regulation. The analysis, using a proprietary world-modeling method, estimate…
A new analysis argues that the infrastructure for secret collusion among large populations of LLM-driven agents, termed 'Private Agent Networks' (PANs), already exists in deployed or developer-facing form, with each nece…
A developer describes how cybersecurity in 2024 relies on layered defenses, visibility, and accepting breaches will happen. In a migration of a 5-petabyte log lake to a cloud-native ELK stack, a random-forest model train…
Norway announced a near-total ban on generative AI tools for primary school pupils aged 6 to 13, effective late August, Prime Minister Jonas Gahr Stoere said at a June 19 press conference, citing risks to foundational ed…
John Jumper, a Nobel Prize-winning scientist who co-led the development of AlphaFold at Google DeepMind, announced he is leaving the company to join Anthropic. The move highlights a broader trend of senior AI researchers…
A blogger announces a return to personal writing in 2026, arguing that human-produced text is essential to prevent AI model collapse as the internet becomes saturated with bot-generated content. The author frames bloggin…
John Jumper, the Google DeepMind vice president and Nobel Prize-winning scientist behind AlphaFold, is leaving the company to join Anthropic, the maker of the Claude AI assistant. His departure signals a major talent shi…
Rep. Ro Khanna, representing Silicon Valley, argues for taxing billionaires like Mark Zuckerberg, citing Meta's $56.31 billion quarterly revenue and simultaneous layoffs of 8,000 workers to fund AI infrastructure. Khanna…
Microsoft researchers detailed an exploit chain called AutoJack that turns an AI browsing agent into a delivery vehicle for remote code execution. The flaw sits in AutoGen Studio, the open-source prototyping interface fo…
Azure AI Foundry, rebranded from Azure AI Studio in late 2024, is a platform of ten distinct capabilities for managing enterprise AI at scale, structured around Hubs for centralized governance and Projects for isolated d…
The Forum of Incident Response and Security Teams (FIRST) concluded its 38th annual conference in Denver, highlighting a record surge in cyber vulnerabilities driven by AI-assisted discovery, with 2026 on track to reach …
Norway will ban generative AI tools in elementary schools starting in late August, with students in grades 1 through 7 prohibited from using AI at all and secondary schools allowing it only under supervision. Prime Minis…
A new Gallup poll reveals that 25% of recently laid-off workers were fully remote, while only 1% lost jobs due to AI. A separate study in Science found that remote work accounts for a third of the decline in mental healt…
Cory Doctorow argued on The Atlantic's podcast that AI companies have lured investors with overhyped claims and accounting gimmicks, predicting the resulting bubble will burst with economic fallout rivaling the pandemic …
Norway will ban generative AI for elementary school students aged 6-13 starting in late August, Prime Minister Jonas Gahr Støre announced. Students aged 14-16 may use AI only under teacher supervision, while older studen…
A researcher mapping the AI safety ecosystem for MATS Research discovered unexpected organizations, including the Human Line Project, which collects stories of AI psychosis, and Impact Academy, which helps professionals …
Confidential Containers, a CNCF Sandbox Project, provides open-source, vendor-neutral technology for protecting sensitive AI models, regulated data, and supply chains using hardware-based confidential computing. The proj…
A cautious optimist argues that AI safety, interpretability, bias, and alignment are as critical as raw capability, warning that the same technology enabling vaccine research could also design pathogens. The author empha…
Google's AI Overview is treating fictional entities from the SCP Foundation horror wiki as real, with no disclaimers. The AI's pattern-matching mistakes the wiki's deadpan bureaucratic style for factual documentation, le…