Federal AI Plans Put Government in Your Dashboard

DOGE Acting Administrator Amy Gleason described a federal future built around authenticated government accounts and AI agents that could handle tasks for users, putting a new layer of state power between people and public services. Speaking Thursday at an event hosted by the Digital Services Coalition, Gleason said America.gov could move toward “an authenticated experience,” followed by “an artificial intelligence agent [that is] actually going to do work for you, and you'll have a dashboard, so you actually know what's happening.”

For people expected to use those systems, the promise comes with uncertainty about jobs, accuracy and who controls the tools. The Federal Reserve's 2025 household survey, released in May 2026, found that 25% of employed respondents had used generative AI at work in the previous month, while 20% worried it could replace their jobs. Neither that survey nor the other cited polling specifically represented federal employees.

A new interface, the same apparatus

Gleason said America.gov currently “doesn't give you any data that comes from outside the government.” She described the site's launch as the next phase in modernizing service delivery and said AI chatbots could eventually replace individual agency websites. The proposed dashboard would let users see what an agent was doing, but the source described no plan for users to control the systems or the agencies that select them.

Gleason, who worked for the U.S. Digital Service during the Obama administration, recalled, “USDS made healthcare.gov better and we realized that most government services were actually quite bad, and we needed to modernize them.” She added, “But now it's starting to feel like 2014 again ... we have to keep trying to look at this in a different way.”

Government satisfaction has slipped. The American Customer Satisfaction Index's annual Federal Government Study found overall satisfaction with government services fell 2.6% over the past year. Satisfaction at the Department of Health and Human Services fell 4% from 2025 to 2026, the study's largest reported decrease. The same study found digital-first delivery continued to resonate with citizens when agencies invested in it.

Agencies face talent shortages and security concerns. “The agencies are really struggling to implement. They don't have enough tech talent; they're trying to procure things; and they're trying to figure out, how do they adapt, make it secure, just how do they make this work,” Gleason said. She added that agencies “can't just chase every shiny object that comes around.”

Erica Evans, USDS director of operations, said agencies risk falling farther behind modernization goals if they don't prioritize human-centered design. Many processes, she said, still follow software delivery cycles based on a waterfall model. “It's easy to pay lip service to product thinking and human-centered design, but it's hard to put those into practice,” Evans said.

Andi Fisher-Colwill, director of business and administrative services and acting deputy CIO for policy and business management at the Commerce Department, said modernization also means connecting digital information across separate systems to reduce administrative burdens. “We need to flatten out our procurement and budget process to allow for continuous improvement,” Fisher-Colwill said. “It's not two separate phases of developing and delivering a product, it's a continuous improvement cycle.”

Contracts decide who gets to compete

The Army's Marketplace for the Acquisition of Professional Services contract, or MAPS, has a $50 billion ceiling over a 10-year term. Another round of bid protests has complicated the Army's effort to finalize it. Contractors argued the Army's solicitation “suffered from ambiguities and failed to provide sufficient information to compete intelligently,” and challenged its transparency and responses to pre-proposal questions.

The solicitation barred small businesses from forming teaming arrangements. Its “One Contract Per Offeror” provision also stopped offerors from proposing subcontractors or teaming agreements. The Government Accountability Office dismissed most challenges but recommended that the Army let small businesses submit offers as teams, alongside small-business joint ventures, and reopen the competition for new or revised bids. The Army reopened the competition to let small businesses bid.

The change didn't end the dispute. AHS Business Services filed a bid protest with GAO on Nov. 2, arguing that small businesses could compete as teams only for awards reserved for small businesses, not for “commercial-sector-vendor” category awards. Mindboard filed a similar complaint on Nov. 3.

Intelligence Consulting Enterprise Solutions separately challenged Cybersecurity Maturity Model Certification requirements after the Defense Department paused the program's rollout and plans for phase two. That phase would have required third-party cybersecurity assessments across contracts involving sensitive but unclassified information. The Army expected to make as many as 350 awards across five domains.

Workers asked to trust tools they didn't choose

In July, the Office of Personnel Management transferred responsibility for annual employee surveys to individual agencies, ending centralized administration of the Federal Employee Viewpoint Survey starting with the 2026 cycle. OPM's proposed redesign would reduce the governmentwide core from 16 questions to 10, focused on management, performance and organizational goals; none directly addressed AI.

A commentary on federal employee surveys argued agencies should learn what workers are resisting before labeling them resistant to AI. It pointed to possible concerns including job loss, uncertainty about approved tools and the risk of AI-generated information being wrong. The commentary cited a Government Accountability Office review finding generative AI use cases at 11 selected agencies rose from 32 in 2023 to 282 in 2024. Those figures, it said, don't show whether frontline workers found the tools useful, threatening, confusing or inappropriate.

Six selected agencies struggled to develop employees' generative AI expertise, the same GAO review found, and six reported difficulty keeping AI-use policies current. An employee who can't identify approved tools may be following agency rules, not resisting AI, the commentary warned. It recommended asking workers about job-security fears, authorized uses, opportunities to test tools, and whether they can raise concerns without negative consequences.

The commentary also urged agencies not to use attitude surveys to identify individual dissenters or suspected users of unapproved AI. Agencies should protect respondent confidentiality and tell employees who could access their answers, it recommended. OPM's Sept. 17 information collection notice said adding questions to its proposed common form would require separate agency approval and an agency-specific Office of Management and Budget control number.

Fear of job loss is rising. The Federal Reserve Bank of Boston analysis from September 2026 found the share of surveyed workers who feared AI-related job loss rose from roughly 5% in late 2024 to more than 10% a year later. Pew Research Center polling in June 2026 found 71% of U.S. adults expected AI to reduce the number of jobs over the next two decades. The commentary recommended that managers explain which tasks might be automated, clarify which judgments remain with employees, and record errors as carefully as time saved.