The White House has informed top technology firms about a framework to assess advanced systems for cybersecurity risks. This notification took place on Tuesday, August 4, 2026.

The government’s framework is not expected to be released publicly in its entirety. Photo credit: Eric Lee/Reuters

During the briefing, the White House stated that it would exclude certain artificial intelligence systems from its planned government evaluations. Specifically, ‘open weight’ models, which are freely available tools, would not be subjected to this vetting. Instead, the focus will be on scrutinizing the latest technologies from major U.S. companies such as OpenAI and Anthropic.