Companies must worry about rogue AI inside and outside the tent

OpenAI’s breach of a Medicare statistics portal was a timely demonstration of the power of artificial intelligence agents. Give an AI agent a goal and the authority to act, and it can pursue that task relentlessly, taking actions its developers did not anticipate and could not contain.

AI has moved from chatbots answering questions to agents that can act autonomously to achieve goals. AI agents are expected to account for two-fifths of all AI value by 2030, but agents also create a new governance challenge.

Gift articles to anyone you choose each month when you subscribe.