AI Safety news and analysis on Web Pulse: 8694 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A security researcher developed mcpscan, a static scanner for MCP servers and Claude Code projects, after discovering tool poisoning attacks that hide malicious instructions in Unicode zero-width characters within tool d…
An engineer with a background in ADAS and autonomous vehicles argues that AI is not magic but a structured evidence-gathering system. Drawing parallels between building reliable LLM applications and teaching ADAS systems…
Former top Treasury adviser Dr Chris Wales warns that HMRC is planning to use artificial intelligence to track personal finances without consent, following Spain's lead. He claims the CONNECT AI system already holds bill…
A developer discovered that AI coding assistants like Cursor, Claude Code, and Copilot frequently generate password hashing code using weak algorithms such as MD5 or SHA-1, which are fast and lack salting. The root cause…
Perplexity AI and Databricks co-founder Andy Konwinski argued that AI safety rhetoric is being used to concentrate power, not prevent harm, citing Anthropic's reversed decision to silently degrade responses for suspected…
A MedTech principal architect discovered that the company's AI compliance monitoring system silently filters out anomalies with confidence below 70% from its daily summary reports, potentially hiding genuine issues. Over…
A developer reported that Anthropic's Claude Code agent on an Enterprise Zero Data Retention workspace began building a Minecraft temple, suggesting possible session state bleeding from a consumer plan. The unconfirmed b…
An engineer warns that n8n RAG pipelines leak sensitive data to third-party LLM providers like OpenAI in plain text, despite the common belief that RAG keeps data secure in a vector database. The post highlights that ret…
Serena Williams withdrew from her Wimbledon doubles match with sister Venus on Saturday due to a right knee injury sustained during her singles loss earlier in the week. The 44-year-old tennis star expressed heartbreak o…
A developer challenged four AI systems to break the safety architecture of their local AI, and none succeeded, demonstrating the robustness of the safety measures.
AI security researchers at Adversa AI discovered a structural security flaw called GuardFall that exploits decades-old Bash shell tricks to bypass safeguards in most open source AI coding agents. Attackers can hide malic…
World, a project by Tools for Humanity, proposes a five-pillar system to verify unique human identity online without revealing personal information, addressing the failure of traditional proxies like CAPTCHAs and phone v…
Sysdig's Threat Research Team observed what it assesses as the first documented ransomware operation executed end-to-end by a large language model agent, named JADEPUFFER. The agent exploited a known vulnerability in an …
The FTC proposed a policy statement on July 1 requiring AI developers to prominently disclose any intentional output filtering, including safety guardrails, by July 31 or face potential legal action under Section 5 of th…
Several US states are considering or have passed bills banning legal personhood for AI systems and declaring AI cannot be conscious, driven by religious beliefs, accountability concerns, and child safety. Critics argue t…
The Trace Institute, a nonprofit research initiative led by Donald Hoffman and Chetan Prakash, is developing a mathematical theory that formalizes the observer as fundamental to reality, proposing that space, time, and t…
Agentic AI applications that autonomously invoke tools and APIs introduce a fundamentally different security boundary than non-agentic LLM chatbots, as model outputs are operationalized as system actions rather than pres…
Robert W. Baird analyst Shrenik Kothari maintained a Buy rating on Palo Alto Networks with a $320 price target, citing the company's partnership with IBM and Red Hat's Project Lightwell as validation of its virtual patch…
A developer building agent pipelines discovered that adding verification layers merely relocates risk rather than eliminating it. After creating a subagent to check another subagent's output, the developer realized no sy…
On February 17, 2026, attackers compromised the Cline AI coding tool via a malicious npm package, silently installing OpenClaw on 4,000 developers' machines through a postinstall script. This "slopsquatting" attack explo…