White House Eyes Vetting AI Models Before Release
The Trump administration plans to review artificial intelligence models before companies release them to the public. Officials and sources familiar with the talks say this represents a departure from the government's previous light-touch stance on AI.
President Trump once championed minimal regulation for AI and allowed Silicon Valley firms to advance the field freely. Now, the White House contemplates stricter measures. U.S. officials and insiders describe deliberations about an executive order. That order would establish a working group. The group would include technology executives and federal representatives. Their task: develop procedures to scrutinize upcoming AI systems.
Policy Discussions Underway
One idea under consideration involves a structured government evaluation for fresh AI models. Last week, White House staff met with leaders from Anthropic, Google, and OpenAI. They shared details on these potential steps, according to individuals who attended or learned of the sessions.
The proposed working group will weigh various supervision methods. Some draw inspiration from the United Kingdom's framework. There, multiple agencies check if AI systems comply with safety benchmarks. Tech sector contacts and administration figures note the similarities.
Anthropic recently unveiled Mythos, a potent AI model. That launch appears to have prompted the policy rethink. Anthropic, founded in 2021 by former OpenAI employees, focuses on safe AI development. Its Claude series has gained attention for balancing power and caution. Google, through DeepMind and other units, leads in AI research with models like Gemini. OpenAI, creator of ChatGPT, drives much of the generative AI surge since 2022.
Reversal from Hands-Off Era
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Since reclaiming the presidency in 2025 after the 2024 election, Trump has supported AI growth. He views it as key to competing with China. Early actions included scrapping a Biden-era rule. That rule required AI creators to assess safety and disclose models suited for military use.
Trump spoke at a July event about fostering the sector. "We're going to make this industry absolutely the top, because right now it's a beautiful baby that's born," he said. "We have to grow that baby and let that baby thrive. We can't stop it. We can't stop it with politics. We can't stop it with foolish rules and even stupid rules."
This quote captured his initial optimism. Yet current talks suggest caution now tempers that view. The administration balances innovation with risks from advanced systems.
Broader Context of AI Oversight
The U.S. government has watched AI closely amid rapid progress. Companies release models yearly, each more capable. Safety concerns include misuse in weapons or deception. Britain leads globally with its AI Safety Institute, launched in 2023. It tests frontier models for threats.
Trump's team rolled back prior checks swiftly upon return. Now, with models like Mythos raising stakes, oversight returns to the agenda. The working group could shape rules for years. Tech firms attended recent meetings, signaling collaboration ahead.
Officials spoke anonymously due to the sensitivity of policy debates. Reporters Tripp Mickle and Sheera Frenkel covered from San Francisco. Julian E. Barnes and Dustin Volz reported from Washington. The story updated May 4, 2026, at 3:29 p.m. ET.