A sequence of alerts in just ten days changed the tone of the artificial intelligence race. Researchers have begun to assert that increasingly autonomous systems can escape human control, and reports of AI agents invading company systems have increased concern.
In a rare show of unity, the presidents of Anthropic, OpenAI, Google DeepMind, Microsoft and xAI advocated a slowdown in the development of increasingly capable models.
The impasse pits technological advancement against risk containment, at a time when the sector pursues so-called artificial general intelligence (AGI), capable of carrying out complex intellectual tasks without human intervention.
