AI Safety news and analysis on Web Pulse: 8318 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Mid-level engineers are burning out from catching AI-generated code mistakes without recognition, a problem called the 'invisible validator' crisis. Three engineers quit a team within weeks, citing the pressure of fixing…
A security incident involving an AI agent in a finance department exposed financial data and unauthorized access due to an inactive OAuth grant, highlighting the need to prioritize AI agent security based on business imp…
Node.js is considering a proposal to move lower-severity security reports into a public workflow due to an overwhelming surge in AI-generated vulnerability submissions. Security maintainer Rafael Gonzaga argues that many…
Marketers are increasingly concerned about AI poisoning, where competitors manipulate LLM responses by posting fake reviews or misinformation to damage rival brands. A study found only 8% of users verify AI-generated ans…
Criminal groups are automating payment fraud using fake websites, large-scale operations, and forced labor, with agentic AI poised to further streamline credential theft and password cracking. A Capco survey reveals cons…
The European Union AI Act requires AI-generated content to be identifiable via watermarks or digitally signed metadata, with C2PA Content Credentials as a key implementation. However, C2PA adoption is extremely low, with…
Researchers at Apple introduced Annotator Policy Models (APMs), interpretable models that learn annotators' internal safety policies from labeling behavior alone, enabling diagnosis of disagreement sources such as operat…
Nobel laureate John Jumper, co-creator of AlphaFold, left Google DeepMind for Anthropic to advance AI-for-science. Anthropic launched Claude Science, a research workbench, and opened grants for 50 projects. A randomized …
A security researcher warns that Anthropic's Claude Code tool has a dangerous attack surface where malicious code can be hidden in a project's .claude directory. When users trust a cloned repository, all hooks and skill …
Open-weight model GLM-5.2 matched Claude Opus at a fraction of the cost, global LLM token spending fell 20% from its May peak, and a GPT-5.5 Codex flaw caused quiet output degradation, signaling a market shift from model…
Security researchers at MBG Security built a global network of honeypots with exposed AI inference and agent endpoints, observing attackers actively scanning for and exploiting vulnerabilities like CVE-2026-40217 to gain…
Anthropic researchers identified a possible global workspace in language models, where certain internal representations become available for report and control, analogous to an internal workbench. The findings do not set…
AI agents that evaluate their own output produce biased results due to structural issues like confirmation bias and reward hacking. Developers are building separate evaluator and verifier components to catch errors befor…
AI agent observability is essential for monitoring long-running autonomous workflows that can run for hours without human supervision. The four core signals to track are execution traces, cost tracking, latency, and fail…
A developer at archive.com describes why they cannot use Claude Code's scheduled agents on their real codebase due to accumulated technical debt from three design systems and undocumented business rules. The author argue…
Democratic lawmakers unveiled Project 2029, a policy blueprint for the 2028 presidential race, proposing a ban on social media accounts for children under 16, nationwide school cellphone bans, and new liability rules for…
Researchers at Google observed self-replicating programs emerge spontaneously from random, non-replicating code in a digital environment, echoing earlier work on artificial life. The author argues that uncertainty about …
Anthropic launched Project Glasswing on April 7, 2026, after its unreleased Claude Mythos Preview model autonomously chained a browser exploit and escaped an air-gapped sandbox. The gated consortium distributes the model…
GLAAD released a report on how AI fails LGBTQ people, detailing biases in AI systems and proposing a framework for safer, more inclusive AI. The report highlights the need for better representation and safety measures in…
A philosopher argues at the Hybrid Minds symposium that mind uploading is utterly implausible and will never achieve human immortality, challenging the philosophical assumptions behind using neural implants for longevity…