AI and Government Services

Legal Risks and Advocacy Under the Federal Government's AI Policies

Mar 03, 2025

The AI Policy Shift

Introduction and Background

The transition to the Trump administration, inaugurated on January 20, 2025, marks a pivotal shift in U.S. AI policy, particularly in the context of government agency deployment. The Biden administration, from 2021 to 2025, had established a comprehensive framework for AI governance, highlighted by the October 2023 Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, which emphasized risk mitigation, equity, and ethical considerations (Administration Actions on AI). This approach included directives for federal agencies to ensure AI systems were safe and responsible, with a focus on addressing biases and protecting civil rights.

In contrast, President Trump's administration, as of March 2, 2025, has adopted a deregulatory stance, as evidenced by his January 23, 2025, executive order revoking Biden's AI policies and promoting innovation to enhance global AI dominance (Fact Sheet: President Donald J. Trump Takes Action to Enhance America’s AI Leadership). This shift, described as moving away from oversight and towards competitiveness, is detailed in analyses like those from Brookings, which note a focus on relaxing agency regulation and competing with China (AI policy directions in the new Trump administration). This change is significant for lawyers, as it may increase the likelihood of legal challenges from clients affected by government AI systems, particularly in areas like public and child welfare and law enforcement.

Understanding the Policy Shift

The Trump administration’s shift contrasts with Biden's "whole-of-government" approach, which included memoranda like OMB M-24-10 and M-24-18 for AI procurement, aiming to ensure responsible use (Key takeaways from the Biden administration executive order on AI). The Trump administration's focus on innovation over regulation could lead to faster AI deployment, but it also raises concerns about increased risks of legal issues, as highlighted in discussions on AI governance challenges (The three challenges of AI regulation).

Potential Legal Issues for Lawyers

Given this policy shift, lawyers are likely to encounter a range of legal issues when clients bring forward concerns about government AI systems. These issues include:

Case Studies and Examples

To provide context, consider the following examples, which illustrate the potential legal challenges:

Jacobin Article: The article "The Fight Against the AI Systems wrecking lives" by Luke Savage, published February 21, 2025, discusses how government AI systems, such as those in welfare, can harm individuals, providing a basis for legal challenges (The fight against the AI systems wrecking lives). The article highlights real-world impacts, such as wrongful benefit denials, which would invoke Due Process concerns if proven.
Child Welfare Algorithm: The Justice Department's scrutiny of the Allegheny Family Screening Tool in Pennsylvania, reported in January 2023, highlights potential bias against families with disabilities, raising concerns about discrimination (Child welfare algorithm faces Justice Department scrutiny). This case, while not yet a lawsuit, indicates a growing area for legal action, particularly under civil rights laws.
Predictive Policing: AI systems in law enforcement, such as those used for predictive policing, have been criticized for reinforcing racial biases, as noted in discussions on AI and civil rights (Accountability in Artificial Intelligence). You may challenge these systems under equal protection principles, seeking to ensure fairness in criminal justice applications.

These examples underscore the practical implications of the policy shift, where reduced oversight could exacerbate existing risks, leading to increased legal scrutiny. This is not good for clients and will undoubtedly spawn more litigation including a flurry of AI related class actions.

Legal Strategies for Advocacy

Constitutional Challenges: Leverage constitutional rights, such as due process under the Fourteenth Amendment, to challenge unfair AI decisions. For instance, Goldberg v. Kelly (1970) provides a precedent for requiring hearings before benefit terminations, which could be extended to AI-driven decisions (Goldberg v. Kelly, 397 U.S. 254).
Administrative Law: Use the Administrative Procedure Act to seek judicial review if agency AI use is arbitrary, capricious, or contrary to law, demanding transparency in decision-making processes, as suggested in legal analyses on AI governance (US federal AI governance: Laws, policies and strategies).
Civil Rights Lawsuits: File claims under federal or state civil rights laws, such as Title VI or VII, if AI systems discriminate against protected classes, ensuring compliance with anti-discrimination statutes, as seen in EEOC actions against AI bias (EEOC Breaks New Ground by Settling First-Ever AI Discrimination Lawsuit).
Data Privacy Claims: Pursue legal action for privacy violations, ensuring government agencies comply with laws like the Privacy Act, particularly in light of AI's data-intensive nature, as discussed in legal issue overviews (Generative AI Legal Issues).
Discovery and Transparency: Use legal discovery to uncover how AI systems make decisions, identifying biases or errors, which can be crucial in building a case, as highlighted in discussions on AI litigation trends (Recent Trends in Generative Artificial Intelligence Litigation in the United States).

This article was not intended to imply that all AI systems are flawed such that their use will inevitably lead to valid claims. But it is undoubtedly the case that the innovation-first philosophy while important in our geopolitical battle with rivals such as China and Russia, will predictably lead to more harms which you will need to have some AI sophistication to address.

As I have spoken and written about for more than 20 years, lawyers do not need to be technology experts. However, you should be aware of what is happening in technology to ensure you don’t miss legal issues in evaluation a client’s claims and/or recognize when the need arises to employ experts in law or technology or both to assist you in doing so.

Key Citations

Legal AI

Discussion about this post