US AI Testing Framework Excludes Open Models

The Trump administration's AI testing framework has been met with criticism for its limited scope and lack of transparency. The framework, which was created in response to President Trump's executive order in June, excludes open models and focuses on closed-source AI models with state-of-the-art capabilities and national security risks. This move has raised concerns about the effectiveness of the initiative and the potential risks associated with closed-source AI models.

The framework sets a 30-day review period for the government to assess new AI models, but the guidelines do not define what 'state-of-the-art' or 'national security risk' means in this context. This lack of clarity has sparked concerns about the framework's ability to effectively address cybersecurity risks.

The Trump administration's decision to exclude open models has also been met with criticism. Open models, which can be downloaded and inspected by anyone, are considered to be more transparent and secure than closed-source models. By excluding open models, the framework may be missing out on the opportunity to assess and address potential cybersecurity risks associated with these models.

The framework's lack of transparency has also raised concerns. The Trump administration has decided not to make the framework details public, sparking criticism from AI companies and experts. This lack of transparency may undermine the credibility of the initiative and make it difficult to assess its effectiveness.

The briefing about the finalized model testing framework was attended by AI companies, including Anthropic, OpenAI, and Google. However, the move has raised concerns about the potential risks associated with closed-source AI models and the need for greater transparency and accountability in the development and deployment of AI technologies.