OpenAI发射Aardvark,Aardvark是一家AI安全代理,使用GPT-5来发现和修补私人贝塔软件的缺陷。 OpenAI launches Aardvark, an AI security agent using GPT-5 to find and fix software flaws in private beta.

OpenAI已推出Aardvark,这是由GPT-5公司建造的AI动力网络安全代理,目的是自主识别和帮助解决软件的脆弱性。 OpenAI has launched Aardvark, an AI-powered cybersecurity agent built with GPT-5, designed to autonomously identify and help fix software vulnerabilities. 目前,该工具是内部开发的,用于支持OpenAI的安保工作,目前正在帮助各组织处理企业和开放源码软件中日益增多的安全缺陷。 Currently in private beta, the tool was developed internally to support OpenAI’s security efforts and is now being offered to help organizations manage the growing volume of security flaws in enterprise and open-source software. Aardvark旨在通过将部分安全研究进程自动化来加快脆弱性检测和补救。 Aardvark aims to accelerate vulnerability detection and remediation by automating parts of the security research process. 虽然该代理商并非公开可访问,但它反映了在高风险领域部署人工智能的更广泛行业趋势,尽管人们对人工智能相关的安全风险感到担忧. While not publicly accessible, the agent reflects a broader industry trend of deploying AI in high-stakes domains despite concerns over AI-related security risks.