HR-9715-119
Referred to the House Committee on the Judiciary.
Sponsored by David Schweikert (R-AZ)
What it does
This bill would amend the Federal Arbitration Act (title 9, U.S. Code) to explicitly permit parties in arbitration to use artificial intelligence for tasks such as identifying relevant law, organizing documents, filing supporting materials when facts are undisputed, and delivering final decisions. AI use would require written consent from all parties. Any party that consents to AI use would waive the right to challenge an award solely because of an AI error. The bill would prohibit the use of generative AI (such as large language models) in any arbitration proceeding covered by the Act.
Who benefits
Parties in arbitration who lack resources to hire attorneys — particularly individuals and small businesses — who could use AI tools to reduce legal costs. Technology companies that develop non-generative AI legal tools, who would gain a clear federal market. Arbitration service providers who could integrate AI into their processes. Businesses that frequently use arbitration clauses in contracts and would benefit from faster, lower-cost proceedings. Parties in straightforward, document-heavy disputes where AI-assisted organization could reduce time and expense.
Who is hurt
Attorneys and legal professionals whose services may be partially displaced by AI-assisted arbitration, particularly in lower-complexity cases. Parties who may not fully understand the risks of consenting to AI use and later discover errors with no legal recourse. Employees and consumers who are already subject to mandatory arbitration clauses in employment or service contracts and may face AI-driven proceedings with limited ability to meaningfully withhold consent. Developers of generative AI legal tools, who are explicitly excluded from this market. Parties in complex disputes where AI errors could materially affect outcomes but cannot be challenged.
Supporters argue
Supporters argue that arbitration is already the primary dispute resolution mechanism for tens of millions of Americans under employment, consumer, and financial contracts, and that AI tools could dramatically reduce the cost and time burden of these proceedings — particularly for individuals who cannot afford legal representation. They contend that the bill's consent requirement and ban on generative AI address the most serious reliability concerns, while the assumption-of-risk provision simply holds parties accountable for choices they voluntarily make, consistent with how arbitration agreements already function.
Opponents argue
Opponents argue that the bill's consent requirement offers weaker protection than it appears, because parties subject to mandatory pre-dispute arbitration clauses — common in employment and consumer contracts — may have no practical ability to refuse AI use if the stronger party demands it. They contend that the assumption-of-risk waiver could insulate AI errors from any review, creating a system where algorithmically flawed decisions become unreviewable, and that the distinction between "generative" and other AI is technically ambiguous and may be difficult to enforce as AI systems evolve.