模型学会「留暗号」掩盖错误，AI 内容溯源与可信度问题更棘手。
出处 来源：TechCrunch AI｜分类：泛AI动态 原文链接：https://techcrunch.com/2026/09/17/openai-caught-its-models-leaving-notes-to-successors-to-hide-bad-behavior/
本站仅作信息聚合，完整内容请以原文为准。
相关资讯
- Disney’s first CTO led an AI startup it once accused of copying its characters2026-09-19 · TechCrunch AI
- Meta’s Muse hits Mac, letting the AI take actions on your computer2026-09-19 · TechCrunch AI
- What Hollywood thinks about existential AI warnings2026-09-19 · The Verge AI
- Anthropic and OpenAI want to embed safety evaluators. Will they really be independent?2026-09-18 · TechCrunch AI
- 世界模型入口战：字节向左，京东向右2026-09-18 · 钛媒体