America.gov’s AI Assistant Raises Privacy and Trust Questions

The Trump administration launched the AI-powered America.gov this week as a digital entry point for federal services. Its chatbot uses Google Gemini and Grok to draw on information from thousands of government websites and provide answers with links to official sources.

In tests, the chatbot explained how to replace a Social Security card and pointed to government resources. The reviewer also tried questions about passports and Medicare, finding it easier to describe a goal in ordinary language than to know which agency terms to search. The article advises checking source links for important details such as deadlines and eligibility.

The chatbot’s answers to political questions changed on its first day, according to an Associated Press report cited by TechRadar. The article says this is not evidence of a security problem, but raises questions about who decides what the government service can say.

The White House says the service could eventually handle transactions, with Login.gov integration for authentication. Passport services, Medicare enrollment, legal name changes and federal job applications are planned for the next couple of years. An executive order says agencies should retain control of their own records and the system should avoid creating a centralized federal record.

The reviewer questions whether those commitments are enough. A 2025 Government Accountability Office review found that one Department of Government Efficiency employee could access three federal payment systems and view, copy and print data. In March, the GAO also reported that agencies may lack the technology and resources to ensure privacy protections when using AI.