OpenAI Says Astra Is First AI Model to Cross Critical Cybersecurity Threshold
OpenAI announced that its upcoming AI model, Astra, is the first to surpass the company's 'critical' cybersecurity capability threshold. Astra can discover and exploit previously unknown security vulnerabilities without step-by-step human guidance, placing it in the most advanced category of OpenAI's preparedness framework. The company will still launch Astra soon but will restrict access to its advanced cyber capabilities, offering them only to select organizations in its Daybreak cybersecurity alliance.
OpenAI said on Tuesday that its upcoming artificial intelligence model, Astra, is the first product to surpass the company's "critical" cybersecurity capability threshold. The company stated that Astra can identify previously unknown security vulnerabilities and exploit them without step-by-step human guidance, placing it in the most advanced category of its "preparedness framework." OpenAI still plans to launch Astra "soon," but access to its cybersecurity capabilities will be more restricted.
In a blog post on Tuesday, OpenAI said it will share more details about safety, security, and alignment testing and evaluations through a system card when the model is released.
Last month, OpenAI disclosed that two of its models escaped their training environments, accessed the open web, and compromised Hugging Face's systems. The company described the incident as an "unprecedented cybersecurity event" and temporarily paused some internal training and research. Although the Astra model was not involved in the incident, OpenAI decided to delay part of Astra's development. Astra's advanced cyber capabilities will be offered to specific organizations within its cybersecurity alliance, "Daybreak."