The Department of Agriculture's $3.9 million contract with Palantir Technologies—potentially worth $13.3 million through 2027—has sparked alarm not because of its size but because it normalizes a government market for software that turns office attendance into an enforcement system. The tool, designed to track return-to-office compliance, could become a prototype for broader federal workplace surveillance.
Meanwhile, the White House has released a mandatory app for government-issued devices that includes official statements, policy announcements, social media feeds, and a “text President Trump” option that pre-populates a praise message. As Government Executive reported, the app also features potential tracking capabilities. Even if the administration frames it as a communications tool, the app raises concerns about political branding and surveillance on devices employees cannot easily ignore.
These two developments may appear separate, but they underscore a common federal blind spot: Washington is adopting tools that can shape, track, or influence the federal workforce before Congress establishes clear rules for their use. This pattern mirrors broader debates about AI surveillance in government, where oversight lags behind implementation.
The White House defends both moves on practical grounds. President Trump's return-to-office directive last year ordered agency heads to end remote work and require full-time in-person attendance, with exemptions. The Government Accountability Office found that many federal headquarters buildings used an estimated 25% or less of capacity during sampled weeks in early 2023. The app, described on the White House website as offering real-time updates and direct access to administration content, addresses communications and management needs.
But good administrative reasons do not erase governance risks. The Palantir contract raises the specter of employee monitoring through mapping, capacity measurement, and analytics. The White House app poses a different risk: mandatory software that exposes workers to official messaging, push alerts, and device-level security questions. Both involve control over the employee environment.
Federal workers experience these technologies directly. A dashboard ranking attendance can become evidence in discipline, relocation, or reduction-in-force decisions. A mandatory app on a government device can serve as a channel for messages employees cannot easily bypass. As Government Executive noted, the Federal Aviation Administration told employees the app would be automatically installed on FAA-issued iPhones and iPads, making the employer the installer.
Congress should treat these stories as part of one oversight category: federal workforce data and device governance. The Privacy Act already governs systems of records, and the Hatch Act aims to ensure nonpartisan administration and protect employees from political coercion. Neither law fully addresses analytics dashboards or forced app installations, but both point to the same principle: agencies should not blur management, monitoring, and political influence. This issue echoes recent debates over FISA renewal and surveillance authority, where partisan divides complicate oversight.
The Palantir system may start with legitimate facilities management, but the same data can support individualized judgments if agencies link attendance, location, supervisory, and HR records. Low office use may reflect telework habits, vacant jobs, obsolete leases, or mission models that no longer require daily in-person work. Data can clarify reality, but it can also flatten it.
The Office of Management and Budget's 2024 memo on AI oversight required agencies to strengthen governance and maintain inventories of AI uses. The National Institute of Standards and Technology's AI Risk Management Framework calls for systems that are valid, reliable, safe, secure, accountable, transparent, explainable, privacy-enhanced, and fair. These standards should apply when software classifies employee behavior or recommends managerial action.
Congress should require agencies to answer key questions before these tools proliferate: What data does the system collect? Who can access individualized records? How long will raw data be retained? Can managers use it for discipline, performance reviews, or office consolidation? Will inspectors general have audit access? Will agencies publish privacy impact assessments and disclose AI functions in public inventories? The fix should not ban public-sector AI or prevent communication, but it must ensure that governance catches up with technology.
