A GitHub page, briefly made public before being taken down, has revealed the Trump administration’s upcoming plan to launch AI.gov, a new federal website designed to bring artificial intelligence tools to U.S. government agencies.
Discovered by researchers at 404 Media, the leaked materials suggest that AI.gov is set to go live on July 4, and will serve as a central hub to help federal departments easily adopt and manage AI technologies.
AI.gov, developed by the General Services Administration’s Technology Transformation Services (TTS), is being positioned as a one-stop platform to streamline how government agencies use AI.
According to the leaked site, the platform includes three main features:
- An AI Chatbot: The site will include a chatbot that can answer questions and help users interact with government systems more easily.
- A Unified AI API: AI.gov will offer a single access point for agencies to use popular AI models from companies like OpenAI, Google, Anthropic, Amazon, Meta, and Cohere. This could help departments use AI without building everything from scratch.
- A dashboard (called CONSOLE): Agencies will also get a live dashboard to track their AI use, monitor performance, and ensure security and compliance.
Who Is Leading the Project?
The project is reportedly led by Thomas Shedd, a former Tesla engineer now working under the Department of Government Efficiency (DOGE), which advocates for reducing government waste and increasing automation.
However, the involvement of DOGE has raised eyebrows. Critics worry the platform could be used not just for digital transformation, but also to justify cuts to federal jobs by replacing human roles with AI systems.
Security and Compliance Concerns
While the platform promises innovation, the leak has also highlighted some serious concerns, especially around data security and regulatory compliance.
For example, not all the AI models listed in the documentation are currently FedRAMP-certified, a necessary requirement for software that handles sensitive government data.
Using uncertified models like Cohere could expose federal systems to cybersecurity risks or legal challenges.
Even more worrying is the way the project was revealed: through a public GitHub repository, seemingly without proper access controls. This accidental exposure has sparked questions about the readiness of the platform and the oversight involved in its rollout.
The AI.gov site is still expected to launch on July 4, but given the public attention the leak has generated, plans could change. Lawmakers, privacy advocates, and tech policy groups are now watching closely to see how the project is handled.
