OpenAI is slowing the development of its upcoming Astra artificial intelligence model after internal evaluations raised concerns that the system may have advanced cybersecurity capabilities that require additional safeguards before release.

The company told Axios it "cannot rule out critical cyber capabilities" in Astra, prompting expanded testing and stricter security measures. The decision marks one of the clearest examples yet of a major AI developer delaying progress on a frontier model because of potential risks associated with its capabilities.

While these abilities could provide significant benefits, researchers have warned that increasingly capable models could also be used to identify vulnerabilities, automate attacks or assist malicious actors.

As a result, OpenAI has increased security requirements around Astra’s evaluation process. The company said it is implementing more isolated testing environments, stronger monitoring systems, and additional controls for applications involving AI agents. The timeline for Astra’s release remains unclear.

The model was not connected to recent cybersecurity incidents involving other AI systems, OpenAI said, but the company determined that further precautions were necessary before moving ahead with a public release.

The decision represents a notable shift in the competitive AI industry, where companies have traditionally raced to deploy increasingly powerful models. Instead of accelerating Astra’s launch, OpenAI is choosing to slow development until it has greater confidence that the system can be deployed safely.

The move also comes as governments and policymakers attempt to establish new frameworks for evaluating advanced AI systems. Officials are exploring how companies should report high-risk models, what standards should trigger additional review, and who should be responsible for assessing potential threats.

Other AI companies have faced similar questions. Developers across the industry have acknowledged that the rapid advancement of AI capabilities has created new challenges around cybersecurity, oversight, and responsible deployment.

Photo: Shutterstock