Origin: Trump administration finalizes secret AI cyber plan
The administration has finished an AI cybersecurity plan and is keeping the details secret. It brought staff from OpenAI, Anthropic, Google, Meta, Nvidia, and other major AI firms to the White House for an overview.
Under the plan, AI developers can voluntarily submit new models up to 30 days before public release. The government will test their hacking‑related capabilities using a classified set of tests and then share the models with federal agencies and selected corporate partners.
The plan is narrowly focused on the cybersecurity features of the most advanced systems, such as Anthropic’s Fable and OpenAI’s ChatGPT 5.6. The White House is not disclosing its testing criteria or which models are covered, creating a pre‑release vetting path for companies while keeping the method and scope out of public view.
Synthesized from 7 sources
AI-synthesized from multiple reported sources
In the app
Read this update in Frameshift
Four stories are free on the web. The full Vault, with every timeline and brief, is in the app.