微软批准用于加沙目标选择的 AI —— 但拒绝透露数据内容
微软承认其 azure 云支持以色列军方 ai，但拒绝检查数据，导致对致命技术的监管无法执行。
following an external investigation by covington & burling, microsoft confirmed that the israeli military used its azure cloud for storage and ai services, including reportedly for ai-powered target selection in gaza. the company announced five new safeguards for conflict zones, including stricter reviews for security agencies and a new 'trusted technology review' channel, following the departure of microsoft israel's country manager alon haimovich. critically, neither microsoft nor the law firm examined the actual contents of the military data, citing customer privacy, which means the operational use of the technology remains opaque despite evidence of terms of use violations in europe.