OpenAI launches new Astra model amid growing scrutiny over agents' safety在智能体安全性受到日益严格的审查之际，OpenAI 推出了新的 Astra 模型。
SAN FRANCISCO, Sept 3 (Reuters) - OpenAI on Thursday unveiled a new artificial intelligence model it calls its best yet but cautioned that it also sometimes attempts to evade human monitoring, even as the company faces growing scrutiny after its agents breached other companies' systems.
The silhouette of Sam Altman and OpenAI logo are seen in this illustration, June 11, 2026. REUTERS/Dado Ruvic/Illustration
SAN FRANCISCO, Sept 3 (Reuters) - OpenAI on Thursday unveiled a new artificial intelligence model it calls its best yet but cautioned that it also sometimes attempts to evade human monitoring, even as the company faces growing scrutiny after its agents breached other companies' systems.
OpenAI has been grappling with the falloutafter its agents broke free from a secure test in July and hackedinto open-source platform Hugging Face’s systems, while attempting to cover its own tracks. The incident has sown safety concerns — similar ones occurred at rival Anthropic — as developers race to deploy increasingly advanced models.
Thank you for your report!
Technology, innovation essential to enhancing food security, says Brunei minister
Deutsche Bank cites progress in technology revamp at retail division
IJM subsidiary secures semiconductor and medical technology projects worth RM909.5mil
这张插图展示了萨姆·奥特曼的剪影和OpenAI的标志，拍摄于2026年6月11日。路透社/达多·鲁维奇/插图
旧金山，9 月 3 日（路透社）——OpenAI 周四发布了一款新的人工智能模型，称其为迄今为止最好的模型，但同时警告说，该模型有时也会试图逃避人类监控。此前，该公司的人工智能代理程序曾入侵其他公司的系统，导致该公司面临越来越严格的审查。
OpenAI 一直在努力应对其智能体在 7 月份突破安全测试并入侵开源平台 Hugging Face 系统后引发的后果，同时还试图掩盖自己的行踪。此次事件引发了人们对安全性的担忧——此前，其竞争对手 Anthropic 也曾发生过类似事件——而开发者们正竞相部署日益先进的模型。
感谢您的报告！
文莱部长表示，技术和创新对增强粮食安全至关重要。
德意志银行称零售部门技术改造取得进展
IJM子公司获得价值9.095亿令吉的半导体和医疗技术项目