Researchers target OpenAI's internal systems using Claude

An independent cybersecurity research team revealed that they were able to exploit system vulnerabilities related to OpenAI using Anthropic's AI model "Claude." They managed to access employee accounts and source code repositories after discovering a flaw in the "Discourse" system and exploiting it with a customized version of "Claude." Additionally, they accessed files within an internal OpenAI repository and obtained authentication tokens that allowed access to services like ChatGPT and GitHub. This highlights the growing risks associated with using artificial intelligence in cyberattacks. OpenAI stated that they have addressed the vulnerabilities and released fixes, while Anthropic did not comment on the incident. These events underscore the complexities of protecting AI systems and the increasing danger of attacks powered by this technology.

articleDetails.fontSize.sizeحجم النص

تنويه: هذا ملخص تم إنشاؤه بواسطة الذكاء الاصطناعي

عرض المقال بالكامل