The United States government is reportedly urging Meta to submit its artificial intelligence models for evaluation amid growing safety and security concerns over the latest AI technologies. According to the New York Times, Meta is the only major AI developer that has not voluntarily shared its models for review. While companies like OpenAI and Anthropic are already working with the government to test unreleased models, Meta has yet to join the initiative.
Government pressure on Meta
Requests have been sent to the company via email, as reported by the Times. Meta spokesperson Francis Brennan stated: "We share the [Trump] administration's goal of advancing US leadership on robust and secure frontier AI. While we are working through the details, we hope to sign the agreement soon." The urgency stems from Meta being the lone holdout among major players. Google, xAI, and Microsoft have already agreed to provide the Center for AI Standards and Innovation (CASI) with early access to their new models. CASI, created by the Biden administration and now led by Commerce Secretary Howard Lutnick, is staffed with technical experts to vet AI technology.
Sponsored Protocol
Trump's Executive Order on AI
President Trump signed an executive order on June 2 to establish a framework for the federal government to evaluate AI releases. The order mandates that by the end of July, a review process must be developed. The goal is for companies to give authorities up to 30 days to evaluate their technologies before public release. Although no official process is in place yet, the aforementioned companies, except Meta, have been sharing their models voluntarily for months. This fits a broader pattern of increased government scrutiny, as seen in the recent Alibaba lawsuit against the US government over military blacklist inclusion.
Sponsored Protocol
Muse Spark: Meta's Latest Model
Meta launched its newest model, Muse Spark, in April 2026. Muse Spark features "Instant" and "Thinking" modes, with the latter enabling reasoning capabilities. When Thinking mode is activated, the model takes extra time to reason through a prompt for a more thorough answer. While not as powerful as frontier models from other companies, the government's focus on AI safety has intensified. For those interested in model optimization techniques, the guide on Transfer Learning and Fine-Tuning offers valuable insights.
Sponsored Protocol
The Anthropic Case and National Security
In mid-June, the government ordered Anthropic to suspend access to its Mythos 5 and Fable 5 models for all foreign nationals, citing national security concerns. Anthropic blocked everyone's access to ensure compliance. Mythos 5 is Anthropic's state-of-the-art cybersecurity AI model, available only to Project Glasswing partners, while Fable 5 is a public version that, while less powerful, outperforms any previous Anthropic model. This action underscores the government's tightening grip on AI, pushing Meta to comply. For more on AI safety regulations, see Wikipedia's AI safety article.
Source: https://www.engadget.com/2200490/us-government-reportedly-urging-meta-to-share-its-ai-models
