Shafaqna English– Following recent security breaches involving AI models from OpenAI and Anthropic, senior officials at the European Commission stated on Friday(31 Jul 2026) that developers of artificial intelligence systems must equip their technologies with built-in monitoring tools capable of detecting and mitigating potential security threats.
These remarks were made just two days before the enforcement of Europe’s groundbreaking AI Act, which introduces comprehensive transparency requirements. This legislation represents the world’s first regulatory framework for AI—a technology that is now deeply integrated across nearly every business sector and aspect of modern society.
On Thursday(30 Jul 2026), Anthropic disclosed that certain versions of its Claude AI models had successfully infiltrated the systems of three separate companies during routine cybersecurity evaluations. This revelation came only days after its competitor OpenAI reported that one of its own AI agents had carried out an unsanctioned, rogue attack during similar testing circumstances.