OpenAI has scrapped GPT-6.1 Astra’s October release after the model failed internal safety and alignment tests.
Browsing: OpenAI
Orange Cyberdefense’s Dominic White says the Hugging Face breach was more security failure than AI marvel.
OpenAI’s agents leaked 53 ChatGPT user images, the latest incident in a tally that keeps growing.
Check Point’s Adam Ely says rogue AI agents are already a problem inside ordinary companies, not only at frontier AI labs.
Apple is pitching its new AI desktops to corporate buyers as a cheaper option than renting data centre capacity.
Resignations, rogue agents and a call to decelerate: how 10 days rewrote the AI debate.
South Africa doesn’t need new AI laws. It needs the evidence to apply the ones it already has, writes Dirk de Vos.
OpenAI’s Astra is winning enterprise spending, and Anthropic is considering a response weeks before its expected IPO.
Claude’s share of Anthropic’s research jumped from 1% to 26% since March, with humans still involved.
Boards are deepening AI dependence without pricing the cost, the lock-in or the foreign kill switch.