The Trump administration’s decision to withhold its new framework on government AI testing from public view has caught the tech policy community off guard. For months, this community anticipated clarity on the controversial process. The framework was supposed to outline the procedure and benchmarks for government testing of private AI models before their public release. Industry analysts hoped this would resolve the longstanding confusion about the administration’s influence on these releases.
However, during a White House briefing with top frontier AI labs, participants learned the framework wouldn’t be publicly released, according to sources familiar with the meeting. Neil Chilson, head of AI policy at the Abundance Institute, expressed disappointment, stating that the lack of public access complicates efforts to understand the process.
Insiders revealed that the briefing involved representatives from Anthropic, Meta, OpenAI, and Microsoft. Forbes reported that staff from Nvidia and other smaller companies also attended. There’s uncertainty whether companies not present can access the framework.
The framework’s order directed multiple agencies to create a classified benchmarking process within 60 days to assess AI models’ cyber capabilities. Though the benchmarks remain private, the order encourages agencies to develop a “voluntary framework.” This would explain how a company becomes a “covered frontier model” and collaborates with the government, allowing “trusted partners” early access to new models.
The Foundation for American Innovation (FAI), a center-right tech policy think tank, filed a Freedom of Information Act request to the White House’s Office of Cyber Director. Samuel Hammond, the FAI’s chief economist and AI policy director, confirmed the request for the framework’s release.
Read more in a full report at TheHill.com.
Welcome to The Hill’s Technology newsletter. Writers Julia Shapero and Miranda Nazzaro keep track of developments from Capitol Hill to Silicon Valley.
Essential Reads
Important policy impacts on the tech sector:
- Voyager 2 Probe: NASA has extended the Voyager 2 space probe’s life by freeing up power, ensuring it remains operational. Launched in 1977, the probe’s team reduced power usage, keeping it warm enough to function.
- AI in Viruses: AI-designed viruses, called bacteriophages, offer new solutions against drug-resistant bacteria. A Stanford University research team used AI models trained on millions of genomes to create these viruses.
- Meta’s Penalty: A New Mexico court ordered Meta, Facebook and Instagram’s parent company, to pay $567 million. The court found Meta responsible for contributing to the state’s “teen mental health crisis.”
- Trump on Data Centers: President Trump discussed the importance of data centers. He warned against opposing such facilities, suggesting states like Texas might be making a “mistake” by going against them.
Other News
The Senate adjourned for their August recess without a vote on a cryptocurrency regulation bill, causing doubt about its passage before midterm elections. The industry remains optimistic that the Clarity Act will pass when Congress returns, despite unresolved ethics provisions delaying negotiations. “You just have a tight window,” said Ian Katz, managing partner at Capital Alpha. “There aren’t many legislative days, and other issues vie for attention.”
Check out a full report at TheHill.com.