AI Safety news and analysis on Web Pulse: 9475 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
OpenAI released a limited preview of the GPT-5.6 series with three tiers—Sol, Terra, and Luna—after the White House requested restricted access. The flagship Sol tier features new reasoning modes and is priced at $5 inpu…
OpenAI released a limited preview of GPT-5.6 on June 26, 2026, introducing a three-model suite (Sol, Terra, Luna) with pricing roughly half that of Anthropic's Claude Fable 5. The launch came less than 24 hours after the…
Anthropic's Claude Fable 5 remains offline on Day 14 after a US Commerce Department export control directive banned the model globally on June 12, 2026, citing a jailbreak that allowed codebase vulnerability scanning. An…
The HALLUCINATE.md Foundation has released an open standard file, HALLUCINATE.md, that instructs AI coding agents not to hallucinate. The minimal specification, consisting of the directive "Do not hallucinate!" in a mark…
OpenAI announced a limited preview of the GPT-5.6 series, including flagship model Sol, balanced model Terra, and fast/affordable model Luna. Terra matches GPT-5.5 performance at half the cost, while Luna offers strong c…
OpenAI is delaying the public release of its GPT-5.6 AI models at the request of the Trump administration, citing cybersecurity concerns. The company will initially share the models with a small set of US government-appr…
A developer documented a full malware analysis lab session, unpacking and analyzing packed binaries, shellcode, and memory forensics. The walkthrough covers static analysis, manual unpacking with a debugger, multi-stage …
A high-severity vulnerability in Amazon Q Developer, tracked as CVE-2026-12957 with a CVSS score of 8.5, allowed malicious Git repositories to execute arbitrary commands and exfiltrate AWS credentials when a developer op…
OpenAI announced three new models—GPT-5.6 Sol, Terra, and Luna—and will initially limit access to a small group of trusted partners at the request of the U.S. government, following a White House request made in consultat…
OpenAI delayed the public launch of GPT-5.6 at the U.S. government's request, limiting initial access to vetted partners shared with authorities. The move reflects Washington's concerns over national security risks from …
Developers concerned about proprietary code leakage and AI hallucinations can build a local code assistant using retrieval-augmented generation (RAG) that indexes their entire repository, answers questions grounded in ac…
Ben Thompson shares his experience with 'vibe coding' an app for organizing his garage, highlighting the fun and potential of AI technology. Apple's Siri AI will not launch in Europe due to the Digital Markets Act, as di…
OpenAI unveiled GPT-5.6, a suite of three models (Sol, Terra, Luna), less than 24 hours after delaying its release at the request of the Trump administration. The company emphasized safety features and cooperation with t…
A new report from the Five Eyes intelligence group warns that AI-fueled cybercrime is increasing, with adversaries using AI to move faster and more effectively. Experts emphasize that defenders must adopt AI to counter t…
OpenAI is releasing a preview of its GPT-5.6 model to a small group of US-government-approved partners, following pressure from the Trump administration to stagger the rollout. The limited release comes amid growing Whit…
Palo Alto Networks completed its acquisition of Koi in April 2026 for an estimated $400 million, folding the startup's supply chain security technology into Cortex XDR and Prisma AIRS. Teams evaluating Koi alternatives n…
Chronic stress from overseeing AI tools is degrading leaders' cognitive function and decision-making, with research showing that high AI oversight loads increase mental fatigue by 12% and information overload by 19%. Neu…
A high-severity flaw in Amazon Q Developer allowed malicious code repositories to silently execute commands on developers' machines and steal AWS credentials. Wiz Research discovered the vulnerability, tracked as CVE-202…
The Trump administration has asked OpenAI to limit the initial release of GPT-5.6 to government-approved partners while federal officials evaluate the model under a new testing framework. The request follows similar inte…
A developer has published a checklist for reviewing AI-agent-generated code before merging, highlighting risky areas such as dependency changes, sensitive paths, and missing tests. The checklist is packaged as a Python C…