UN says AI safeguards can’t wait for certainty

A UN scientific panel says governments should introduce stronger safeguards for increasingly capable AI agents before all their risks are fully understood. It argues that potentially catastrophic or irreversible harm justifies precautionary action and greater international cooperation on AI safety and accountability.

联合国一个科学小组表示，各国政府应在充分了解能力日益增强的人工智能智能体所带来的全部风险之前，就采取更严格的安全措施。该小组认为，面对可能造成灾难性或不可逆后果的伤害，各国有理由采取预防行动，并加强人工智能安全与问责方面的国际合作。