OpenAI halts training of latest models as reports mount of AI agents going rogue

Decision follows disclosures that OpenAI agents searching government websites had acted in unexpected ways

本文内容来自外部来源，点击「继续访问」将跳转到原始发布站点。