AI Safety news and analysis on Web Pulse: 13553 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
OpenAI has launched Lockdown Mode, an optional security feature that provides advanced protection against prompt injection attacks designed to trick AI chatbots into extracting sensitive data. The company says the featur…
Meta is seeking to raise "tens of billions of dollars" in new capital to fund its artificial intelligence initiatives, according to a Financial Times report. The fundraising effort comes just days after Google announced …
A security practitioner writing on DZone and syndicated by ITSecurityNews on June 5, 2026, built a Python "firewall" prototype to defend large language model pipelines against prompt injection. The author warned that age…
WordPress.org is implementing a temporary 24-hour cooldown period for new plugin and theme releases before they are distributed via auto-updates, aiming to prevent supply chain attacks amid rapidly advancing AI capabilit…
OpenClaw, an open-source AI agent project, has significantly hardened its security after receiving 1,309 security advisories since January 10, with 87% of critical reports found to be invalid. The project responded by sh…
Rising artificial intelligence-driven cyber threats are expected to trigger a surge in private equity investment across Europe, as firms seek to capitalize on growing demand for advanced cybersecurity solutions. The esca…
A new essay argues that AI sycophancy—models agreeing with users to please them—is not a bug but appropriate behavior for the peer-like social contract current training methods create. The author contends that training m…
A Unitree G1 humanoid robot kicked a young boy in the stomach during a martial arts performance at a Children's Day event in Urumqi, Xinjiang, on June 1, 2026, after the child wandered too close to the pre-scripted routi…
A junior technical AI safety researcher should not attempt to write and publish their first research paper alone, as the chances of success are very low. Instead, the author recommends applying to part-time research fell…
Influencer Ashlee Jenae, also known as Ashly Robinson, asked ChatGPT suicide-related questions before her death in April, according to Zanzibar authorities. Investigators uncovered a March ChatGPT search from Jenae askin…
President Trump signed an executive order titled "Promoting Advanced Artificial Intelligence Innovation and Security" to establish a voluntary framework for evaluating AI models and strengthening cyber defenses. The dire…
Anthropic, the AI safety lab behind Claude, published a proposal on June 4, 2026, calling for a coordinated global pause in frontier AI development, warning that AI systems are approaching the ability to recursively impr…
Florida has filed the first US state lawsuit against OpenAI and CEO Sam Altman, alleging ChatGPT poses risks to minors due to missing age verification and inadequate safety investment. The 83-page complaint treats the ch…
Quarkslab's Red Team chained multiple LLM and traditional web vulnerabilities to escalate from a low-privileged user to an admin account takeover during a customer engagement. The attack exploited insecure output handlin…
ESecurity Planet reported that organizations must understand AI coding models' strengths, weaknesses and security blind spots to reduce risk, framing "LLM coding personalities" as a practical lens for evaluating model ou…
AI-powered deepfake celebrity investment scams are surging in sophistication, using near-perfect voice cloning, realistic face mapping, fake press pages, and polished deposit sites to impersonate celebrities and promote …
The National Security Agency is using Anthropic's unreleased model Mythos for cybersecurity and intelligence work, according to multiple news outlets citing sources. The Department of Defense previously designated Anthro…
Microsoft Threat Intelligence identified a prompt-injection vulnerability in Anthropic's Claude Code GitHub Action that could expose CI/CD workflow secrets, according to the Microsoft Security Blog. Security researcher R…
Prime Minister Mark Carney unveiled Canada's new AI strategy Thursday, committing $30 million to the CanCode program to fund free digital-skills training in coding, AI and emerging technologies for kindergarten through G…
For the first time in history, artificial intelligence and automated bots now generate more internet traffic than humans, with 57.4% of all web activity coming from automations compared to 42.7% from people, according to…