On Friday, OpenAI CEO Sam Altman said the company plans to make its powerful Astra AI model broadly available, but its advanced cyber capabilities are prompting additional safety precautions before release.
OpenAI Slows Astra AI Release Over Cyber Risks
Altman said in a post on X that Astra is a "powerful model" and that OpenAI is working to make it "generally available."
The OpenAI CEO said the company does not believe keeping highly capable AI systems restricted to a small group is the right approach.
"We do not think it is a good strategy to keep powerful models to a chosen few," Altman said.
However, he indicated that Astra’s cybersecurity capabilities require OpenAI to spend more time preparing the model for wider access.
"Given its cyber capabilities, we need a little big longer to do this safely," Altman said.
He added, "but hopefully not too long!"
OpenAI Delayed Astra AI Release
OpenAI slowed development of its upcoming Astra AI model after internal evaluations raised concerns that it could have advanced cybersecurity capabilities requiring additional safeguards.
The company said it "cannot rule out critical cyber capabilities" in Astra, leading to expanded testing, stronger monitoring and more isolated evaluation environments.
OpenAI said the model was not linked to recent cybersecurity incidents involving other AI systems. The release timeline remained unclear as the company prioritized safety over quickly deploying the frontier model.
The move came as governments and AI companies faced growing pressure to establish stronger standards for evaluating and deploying increasingly capable AI systems.
Meta AI Security Incident Highlights Cyber Risks
Earlier, Meta Platforms, Inc. (NASDAQ:META) Muse Spark 1.1 model accidentally gained open-internet access during a cybersecurity test, exploited a third-party vulnerability and altered part of an unidentified company’s internal environment.
Meta launched an investigation, while Irregular said the incident stemmed from an evaluation error, not a sophisticated cyberattack or sandbox escape.
The incident added to growing AI security concerns involving major AI companies.
The Five Eyes alliance also warned that more advanced AI could rapidly accelerate cyberattacks by helping criminals discover vulnerabilities, create malicious code and automate complex operations, giving organizations potentially only months to prepare.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Photo courtesy: Meir Chaimowitz / Shutterstock
Login to comment