One Nation, Under Artificial Intelligence: An Overview

Summary:
The U.S. federal government is increasingly implementing artificial intelligence (AI) to enhance its operations. Despite this upward trend, there has been limited discussion about the associated human rights concerns. The Government Accountability Office (GAO) and other studies reveal a significant rise in AI use among federal agencies, highlighting its potential benefits and the need for oversight.

Key Takeaways

The use of generative AI in U.S. federal agencies has surged, with reported applications climbing from 571 in 2023 to over 1,100 in 2024.
Leading departments utilizing AI include Health and Human Services, Veterans Affairs, Homeland Security, and the Interior, collectively representing half of public AI applications.
AI is primarily used for improving operational efficiency in various sectors such as health, public services, and cybersecurity.
State and local governments are also exploring AI, aiming to enhance budget efficiency and governance.
There are concerns about how AI could potentially undermine public rights and safety, prompting calls for more transparency and regulation.

The federal government is now deploying artificial intelligence operations, while acknowledging some rights are at risk.

Increasing AI Adoption in Federal Agencies

The rollout of artificial intelligence in the public sector is progressing rapidly, with U.S. federal agencies embracing AI to improve operational processes. Recent studies have uncovered a substantial increase in AI implementation, despite a lack of widespread public discourse regarding its implications for human rights.

A report from the Government Accountability Office revealed a dramatic increase—nine times the growth in generative AI use from 2023 to 2024. The report analyzed 11 federal agencies, finding that reported AI applications nearly doubled from 571 to 1,110. Additionally, the Chief Information Officer’s Council published AI in Action: 2024 Federal AI Use Case Inventory Findings, confirming that AI utilization in federal agencies more than doubled since 2023, with more than 1,700 documented applications spanning internal support and public services. Agencies noted gains in operational efficiency and mission execution as key reasons for this uptick.

Key Players in the AI Rollout

Several federal agencies are at the forefront of AI integration, including the Department of Health and Human Services, Veterans Affairs, Homeland Security, and the Interior—these departments account for 50% of reported AI uses.

The Department of Homeland Security has been particularly proactive, leveraging AI and machine learning to enhance capabilities in cybersecurity, border security, and disaster response. AI tools have allowed the DHS to strengthen investigative leads, support local governments with hazard mitigation, and facilitate training. They maintain transparency by publishing a comprehensive list of AI tools in their inventory.

The Centers for Disease Control and Prevention (CDC) has utilized AI to expedite investigations into multi-state foodborne disease outbreaks. Furthermore, the Veterans Benefits Administration employs AI to protect veterans' benefit payments by detecting fraudulent changes in direct deposits. The Department of Health and Human Services has employed AI to analyze publications and identify outbreaks in previously polio-free areas, while the Social Security Administration utilizes it to support adjudicators in disability programs.

In an effort to broaden its AI capabilities, the U.S. General Services Administration has added leading American AI products, such as Anthropic’s Claude, Google’s Gemini, and OpenAI’s ChatGPT, to its inventory.

Diverse Applications of AI in Governance

The GAO's report, Artificial Intelligence: Agencies Have Begun Implementation, outlines how AI is utilized for analyzing drone imagery and various datasets across federal agencies. Among the 23 agencies reviewed, 20 reported around 1,200 current and projected uses of AI, which specifically includes analyzing camera data from drones and radar to monitor border activities as well as identifying scientific specimens for planetary rovers. NASA and the Department of Commerce exhibited the highest number of AI applications.

In addition, the Administrative Conference of the United States has offered recommendations for AI's use in governance, highlighting its functionality in regulatory enforcement, adjudication, and automated legal guidance.

Federal agencies have reported benefits from AI in the form of improved anomaly detection, streamlined processes, and enhanced decision-making. Currently, most AI applications fall within administrative and IT functions, followed by health and medical usages. About 13% of use cases pertain to health and medical applications, and roughly 9% assist in processing government benefits, including Medicare, Medicaid, Social Security, and unemployment services. Notably, 46% of the AI applications within federal agencies are deemed "mission-enabling," improving various administrative functions.

Federal agencies are actively managing the application of AI across various sectors such as law enforcement, education, transportation, and environmental sciences, collaborating with both internal experts and corporate partners. In-house AI development constitutes 50% of the instances, with over 40% relying on custom-developed code that is publicly accessible.

State and Local Government AI Initiatives

State and local governments are also exploring AI to enhance their governance efficiency. A report by Deloitte indicates that these entities may be turning to AI to alleviate budget constraints. In Georgia, an "AI Accountability Act" was proposed to establish a board overseeing AI mandates, requiring government entities to formulate AI usage plans, data privacy measures, and specify the necessity for human oversight.

Similarly, a bill in Montana aimed to restrict AI's role in state governance, advocating for transparency about AI applications and mandating human review for specific recommendations. However, both these legislative efforts stalled in committee.

The Human Rights Consideration

Approximately 13% of federal AI applications carry the potential to infringe upon public rights or safety, as outlined in the OMB Memorandum M-24-10. Agencies are encouraged to implement robust safeguards, including assessments and monitoring to mitigate the risks of algorithmic bias and discrimination.

By December 2024, agencies had conducted AI Impact Assessments for over 80% of AI applications affecting rights or safety, with more than 70% of these cases undergoing independent evaluations. Notably, nearly half of the rights-impacting AI applications within government feature mechanisms for appeals or procedures to contest AI-generated outcomes.

As AI is increasingly embedded in governmental processes, it is crucial to engage in discussions addressing its implications for daily life and rights. Public discourse surrounding the ethical use of technology in governance should be prioritized, ensuring transparency and accountability as AI continues to evolve.