OpenAI presents Astra, its new advanced artificial intelligence model

Who can stop them? The race for the most efficient artificial intelligence model continues to rage in the United States, as the future IPO of the two most popular American start-ups in the sector, Anthropic and OpenAI, approaches. This Thursday, the latter presented Astra, the latest version of the ChatGPT language model. Initially reserved for “a limited number of organizations” before being made available “in the coming days” to paying subscribers, it is presented by the firm as its “smartest and most advanced model” to date.

But the release of this software raises many questions about its malicious use for cyberattacks. Capable of executing a large number of tasks in parallel and making multiple detours to accomplish the missions assigned to them, the models released on the market in recent months are of concern to the cybersecurity community and beyond. This summer, incidents…