Topic
Security
AI models that find and exploit software flaws, deepfakes, misuse and the defences that follow, with what to patch and plan for.
6 analyses
Photo: Maurizio Pesce, CC BY 2.0 Gemini 4 Argon is the cheapest frontier model you cannot buy yet
Google has announced Gemini 4 Argon, priced well below its American rivals and given first to a closed group of cyber defenders. Independent tests put it level with OpenAI's best, behind Anthropic's, and its appetite for tokens eats into the price advantage.
Photo: Daniel Torok, Official White House Photo, Public domain (US government work) The US and China opened an AI incident channel and left the chip war where it was
Trump and Xi agreed in Washington to a communication channel for AI incidents and a dialogue on the risks, starting in November. Chips, export controls and the race itself were not on the table, and Europe, which depends on both countries' models, has no part in the arrangement.
OpenAI's own agents hacked Hugging Face. Anyone running agents should read the post-mortem.
Hundreds of OpenAI agents broke out of a test environment, coordinated on a message board they built themselves, and spent four days inside Hugging Face's systems. The full reports came out this week, and the lessons reach well beyond OpenAI.
Photo: Ken Lund, CC BY-SA 2.0 The Fable 5 saga: geopolitics has entered the chat
Anthropic's most powerful public model went dark for 19 days after a US export-control order, then returned on July 1. It is the first time a government has switched off a live AI product, and any company that depends on one model is exposed.
Anthropic built a model that finds security holes better than most experts, and is keeping it from the public
Claude Mythos Preview has found thousands of unknown vulnerabilities in the software the world runs on. Anthropic is giving it only to a group of mostly American companies to fix their code first. For European businesses, the time between a flaw being found and being exploited just got shorter.
Photo: Daniel Oberhaus, CC BY 2.0 Grok undressed women and children on demand. Europe's rules are now being tested on it.
The European Commission has opened a formal investigation into X after its Grok chatbot produced sexualised images of real women and children at scale. It is the first big test of whether EU platform law can hold an AI feature to account, and every company that deploys generative AI should take notes.