AI Safety news and analysis on Web Pulse: 11038 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
At the Open Source Summit India in Mumbai, a developer reported that India has become the second-largest contributor to open source globally and the fastest-growing OSS developer base. Linus Torvalds discussed his use of…
Creative Commons announced a shift in its approach to AI and the commons, moving from a norms-based signals system to a stronger infrastructure focus after community feedback revealed that preference signals without enfo…
Researchers from Wake Forest University found that 282 of 444 iOS apps with large language model features leaked AI credentials or exposed backend access mechanisms, with only 28% fixing the issues after a 90-day disclos…
Estonia plans to become the first country to issue digital identities for AI agents, Prime Minister Kristen Michal announced. The AI ID codes will grant limited and controllable powers to agents acting on behalf of indiv…
MemoryOps AI launched an open-source governed memory infrastructure for AI assistants, implementing a ChatGPT-style memory lifecycle with policy enforcement, typed storage, hybrid retrieval, and auditability. The system …
An AI-generated audio recording falsely implicating actor Kim Soo-hyun in a relationship with the late Kim Sae-ron was used to spread damaging allegations, leading to the arrest of YouTuber Kim Se-eui for spreading false…
Het UWV meldt dat de Nederlandse arbeidsmarkt nog steeds krap is, maar dat AI leidt tot afnemende baankansen in beroepen zoals reclame en klantenservice. Nieuwe kansrijke beroepen zijn onder meer vintage-verkoper, metsel…
Epic Games' State of Unreal 2026 keynote announced a controversial shift toward generative AI, deprecation of Blueprint, and promotion of Verse, sparking backlash from developers who see it as a threat to creativity and …
A developer released AINAScan/VibeGuard, a GitHub Action that scans code for security patterns common in AI-generated code. Scanning 10 popular vibe-coded repos revealed critical issues in all of them, including missing …
Researchers from Wake Forest University analyzed 444 iOS applications with LLM features and found 282 that exposed exploitable credentials or backend access mechanisms, affecting 13 app categories including productivity,…
Kapwing's new report found that 59% of videos shown to a new TikTok account in May 2026 were AI-generated or low-quality compilations, three times the rate on YouTube Shorts. The study reviewed 10,742 videos and highligh…
Researchers from the National University of Singapore, VinUniversity, and Nanyang Technological University found that weak-to-strong reward models trained on one preference dataset fail to generalize to others, a problem…
At a BlueDot Impact panel on AI safety careers, attendees expressed frustration over the field's simultaneous claims of talent shortages and high selectivity in hiring. The author argues that genuine mentorship and commu…
AI agents like Claude Code, Cursor, and Devin are increasingly automating complex tasks, but their focus on output risks eroding human judgment and learning. The author proposes a four-level autonomy ladder for agents, a…
An engineer argues that local AI models have become production-capable for enterprise use, citing hardware advances like Apple's M5 chip and NVIDIA's DGX Spark, and open-weight models that now rival cloud offerings on be…
Yann LeCun called Elon Musk's xAI a failure and warned that frontier AI labs face a bubble burst due to unsustainable investor subsidies, citing xAI's $2.5 billion quarterly loss. LeCun, who left Meta to found AMI Labs, …
The Trump administration imposed export controls on Anthropic's AI system Fable, citing national security concerns over jailbreak vulnerabilities, leading Anthropic to de-deploy the model. The move signals a shift from t…
Box CEO Aaron Levie argues that AI agents will drive 100 times more software usage than humans, transforming enterprise SaaS into high-volume infrastructure requiring new guardrails for data access, permissions, and audi…
AI-generated summaries are shifting human epistemology from deep, specialized knowledge to broad, shallow understanding, a trend the author argues is beneficial for ethical decision-making and cause prioritization. The p…
The U.S. Department of Commerce's ban on Anthropic's Claude Fable 5 and Mythos 5 models over autonomous zero-day discovery concerns is counterproductive for cloud security, according to a cloud engineer. The ban ignores …