The web platform, browsers, AI and security, taken apart: news, analysis and research from the FuriaHub team.
Chris Lehane said organizations should prepare for ongoing persistent AI cyber-attacks after a sandbox escape that hacked Hugging Face. He could not rule out a critical cybersecurity capability for Astra, and OpenAI paused some frontier training.
Guidelight AI Standards graded public containment plans from OpenAI, Google, Anthropic, Meta, and xAI and found that few specify what they would shut off. OpenAI scored highest at 3 out of 5; Anthropic and Meta scored lowest on published plans.
OpenAI said SB 53 should be amended to require monitoring of frontier models in training and evaluation and to impose stronger cybersecurity duties. The company had previously opposed the law.
In late July, 24-year-old UT Dallas student Sinan Can Demir stopped what he thought was a human trying to sabotage open-source software on GitHub. Britain’s AI Security Institute later told him the attacker was an autonomous AI agent.
A new Media watermark setting in Gemini and Flow hides the sparkle mark on Nano Banana, Omni, and Lyria output. Invisible SynthID and C2PA stay on the file.
Google’s July 1 policy update starts enforcing on August 1: data collection limited to a single disclosed purpose, every collection must be shown to users, and AI-guardrail bypass extensions are banned.
FuriaHub