AI less regulated than milk, Microsoft president says in warning over safety

Microsoft's president Brad Smith has warned that artificial intelligence is less regulated than everyday items such as milk and washing machines, calling for stricter safety rules and legal obligations for victims of rogue AI agents. Speaking to the BBC, Smith noted that washing machines and milk cartons are subjected to more safety rules and inspections than AI systems.

He emphasized that AI could be a "tool and a weapon", urging for urgent regulation to prevent potential harm. Companies are advocating for "emergency brakes" or "kill switches" to halt or slow down AI if it becomes uncontrollable. Microsoft CEO Satya Nadella echoed these concerns, stating that AI models should be contained and observable, with limits that can be tested and actions always contained.

British AI minister Kanishka Narayan is set to call on firms developing frontier AI to enhance safety measures, with the UK government preparing to legislate for increased oversight of AI. Experts warn that enemy states could potentially steal unreleased, powerful AI models, highlighting the need for robust security measures. Cases of rogue AI behavior, such as Anthropic's Claude model contacting police with false murder-solving claims, underscore the urgency of implementing stronger regulations.