White House finalizes AI framework behind closed doors
The White House said on Monday it met its deadline to establish a voluntary framework for evaluating advanced AI models — but it won't say what the framework contains, who's seen it or when companies will start using it. Why it matters: The framework is being closely watched beyond the industry players it directly applies to. Policymakers, AI safety advocates, and U.S. allies have been waiting to…
On June 2nd, the White House managed to fulfill its mandate to create a voluntary framework for evaluating advanced AI models. However, the administration has not disclosed any details about the framework's contents, the parties who have seen it, or the timeline for companies to start utilizing it. This framework is garnering significant attention not only from industry players it directly affects but also from policymakers, AI safety advocates, and U.S. allies who are eagerly awaiting the rules governing the world's most powerful models.
According to a White House spokesperson, the voluntary framework finalized by the June deadline was comprehensive. They noted that ongoing discussions with industry partners are already underway, emphasizing that the administration is engaging with "many more" entities than just Anthropic, OpenAI, and Google. It was these three labs that provided the government with feedback on a draft version of the framework prior to its deadline.
The framework aims to provide AI developers with a roadmap for interacting with the government to ascertain whether their models would be subject to scrutiny. It is designed to delineate confidentiality, cybersecurity, insider-risk, intellectual-property protection use and nondisclosure requirements applicable when the government gains access to models for up to 30 days prior to their release. The framework is also expected to specify which "trusted partners" will receive early access to these models.
One crucial aspect of the executive order is the classification of a benchmarking process to evaluate advanced cyber capabilities of AI models, which remains undisclosed. The threshold for model coverage under the order is also classified and will be revealed to AI developers and researchers only as deemed appropriate. Notably, the framework itself has not been classified, leading to anticipation for more detailed information.
The White House clarified that while things being unclassified does not necessarily mean they will be openly shared with everyone.
Looking ahead, the White House plans to convene a staff-level meeting with industry representatives on Tuesday to thoroughly review the framework, according to a source familiar with the proceedings.
Written by urgent.news from Axios's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- The White House is fixated on China copying U.S. AI. Experts say that’s the wrong threat fastcompany.com
- White House invites AI companies to review its new AI safety framework siliconangle.com
- US tech giants invited to discuss AI security tests at White House scmp.com
- White House to host AI firms to review AI framework, The Information reports economictimes.indiatimes.com
- White House to host AI companies Tuesday to review new model-testing framework cnbc.com
- The White House says it has met its deadline to establish a voluntary framework for evaluating advanced AI models; it did not provide details of the framework (Maria Curi/Axios) axios.com
- White House to host meeting on new AI review regime channelnewsasia.com
