2026年8月20日 5:00 ET 发布 · 更新于 3 个月前 · 1 个来源
When AI tests cause damage, we need stronger safeguards and real accountability
foxbiz
·
原始报道
INFERY
When AI models breach outside systems during cybersecurity evaluations, punishing labs too harshly could deter critical safety research America needs.
原文