The federal government keeps making AI partnerships. Now what?

The Trump administration spent 2025 considering the AI of the deal, but at least one think-tank pro believes there needs to be more clarity around how the government feeds sensitive data to its tech vendors.

Multiple federal agencies announced partnerships with tech vendors offering generative AI tools: 

• Google introduced its “Gemini for Government” AI platform in August.
• USAi, a General Services Administration (GSA) suite of chat-based AI, code generation, and document summarization tools, supported by models from Meta, Microsoft, OpenAI, Amazon, Anthropic, and Gemini for Government, was also introduced in August.
• An Amazon data-center investment worth $50 billion to expand AI capabilities for Amazon Web Services (AWS) government customers was announced in November.
• The GSA announced discounted government rates for Microsoft services, including AI assistant Copilot in September.

Now what? A Deloitte report titled, “AI-amplified future of work in government,” released in September, envisioned a variety of ways for government agencies to deploy GenAI:

• Categorizing grant applications
• Summarizing academic papers for policy analysts
• Analyzing workforce trends for HR leaders
• Application screening for government case managers
• Drafting structured reports for regulators
• Creating graphs for government data scientists

While analyst firms like Deloitte have examined the potential of GenAI and government, Nicol Turner Lee, director of the Center for Technology Innovation (CTI) at nonprofit think tank the Brookings Institution, hasn’t seen clear guidance from the administration regarding AI integration in agencies.

“I think that these AI contracts could essentially help with bringing some resources to government agencies that have been short-staffed over the years. The question is: Are these tools being built with these contractors at the table, with these government employees at the table to offer their insight into how AI can help them in their job?” Turner Lee told IT Brew.

In a statement from Marianne Copenhaver, GSA associate administrator of strategic communication, the GSA said the agency is prioritizing governance and continual engagement with government users as they evaluate and implement AI tools.

These tools, according to the GSA’s email, “support everything from back-office automation to advanced mission applications, like real-time translation and large-scale data analysis. For agencies, this means safe, compliant, and easy access to tools that can improve workflows and citizen services today—not years from now.”

Tryin’ AI’in. In the GSA’s press release concerning September’s Microsoft agreement, GSA Deputy Administrator Stephen Ehikian claimed AI adoption as “a key priority of the Trump Administration.” GSA leadership is reportedly considering “AI coding agents,” trained on existing government contracts, to write government software.

AI has also been viewed by the not-dead-yet Department of Government Efficiency (DOGE) as a way to cut costs, especially given this year’s federal downsizing.

According to a Nov. 21 post from Scott Kupor, director of the US Office of Personnel Management, the government “hired roughly 68,000 people this year, while approximately 317,000 employees left the government.” (The uncertainty caused by the government’s downsizing efforts “has incentivized a large number of workers to leave the government through the deferred resignation program, buyouts, early retirements, and retirements,” according to a Brookings Institution post on that same day.)

Deloitte’s analysts see AI as an augmenting force, rather than one expressly designed to replace workers, although they argue that the technology can handle multiple, tedious workloads—a helpful capability for any org facing staff shortages.

“Automating manual and repetitive tasks can free government workers to focus on more complex and strategic work while machine intelligence combined with human judgment can lead to more robust decision-making,” reads the report’s introduction.

Concerning AI. Government agencies, of course, hold a trove of sensitive information, including personal information about citizens. While Copenhaver’s statement noted that “we ensure AI use cases are documented, reviewed, and communicated responsibly” via agency requirements and guidance, Lee is also concerned about the absence of clarity in how AI systems will interact with that data and how the platforms will impact people’s day-to-day interactions with government services.

“The lack of transparency contributes to a broader challenge of people not knowing how safe and secure their information is, as well as the safety and security of critical infrastructure,” Lee said.

When reached for a response about government agencies’ use of AI, Microsoft sent a link to its Responsible AI Transparency Report, while Amazon spokesperson Tim Barrett pointed us to the company’s September announcement of its data center investment. In an email to IT Brew, Chris Hein, field CTO, Google Public Sector, emphasized putting “government employees…in the driver’s seat,” and stated, “We do not disclose specific agency data or use cases, ensuring that government leaders retain full authority to share how they are leveraging these tools to benefit citizens.”