早期 AI 事故多因糟糕的安全实践，暴露行业对齐短板
jessi_cata · x · 2026-09-30
Commenting on claims that OpenAI didn't know it needed to monitor AI agents, Jessi Cata argues that regardless of how hard securing/aligning a given capability level is, bad AI incidents are themselves evidence of bad security and alignment practices — so it's unsurprising the first bad incidents involve bad practices.
「安全」频道最新
- 因 Hugging Face 泄露事件，OpenAI 遭用户起诉 — wiredmagazine · 2026-09-30
- 特朗普重申对 AI 行业轻监管,称行业自律已足够 — pstAsiatech · 2026-09-30
- 学者呼吁 FTC 关注 AI 产品营销,强调需保护未成年人 — LuizaJarovsky · 2026-09-30
- Anthropic 红队警告:一款可下载的中国模型已能自主制作攻击工具 — ross2000 · 2026-09-30
- 美国众议院议长 Johnson:希望 AI 监管「自愿」,国会立法停滞 — sourdub · 2026-09-30
- 博主复盘 OpenAI 澳洲政府服务器被黑事件的来龙去脉 — luisdans · 2026-09-30