OpenAI Briefs Governments On New Cybersecurity-focused AI Tool

OpenAI has begun briefing U.S. federal agencies, state governments and members of the Five Eyes on its new cybersecurity-focused AI model, GPT-5.4-Cyber, according to a report by Axios.

The report said the company recently hosted a closed-door event in Washington for around 50 government cyber officials to demonstrate the tool’s capabilities. Access to the model is being controlled through a vetting system under its Trusted Access programme.

Exclusive: OpenAI briefs feds and Five Eyes on new cyber product https://t.co/Y6B9K4u1Za

— Axios (@axios) April 22, 2026

According to the report, OpenAI is offering a dual-track rollout. One version includes safeguards for broader use, while a more advanced version is reserved for trusted defenders. Officials said the tool could help identify vulnerabilities in outdated government systems.

The report also noted that rival Anthropic is testing a similar model with limited access due to security concerns.

OpenAI plans to expand briefings to allied governments to coordinate cybersecurity efforts.

Also Read:

Why Democrats Could Gain Seats After Virginia Vote

Voters in Virginia approved a new congressional map that could give Democrats up to four additional House seats, according to a report by CNN. The map, passed through a statewide referendum, is expected to favor Democrats in 10 of the state’s 11 districts. The report said the move is

tippinsightsTIPP Staff