Intelligent Arbitration Act of 2026
- Sponsor
- Rep. Schweikert, David [R-AZ-1]
- Committees
- Judiciary Committee (primary)
- Last action
- Jul 15, 2026
Bottom line
This bill legalizes and regulates the use of non-generative artificial intelligence in arbitration, requiring party consent and limiting appeals based on AI errors, aiming to modernize dispute resolution processes.
What it actually does
This bill amends Chapter 1 of title 9, United States Code, to explicitly permit parties in arbitration to use non-generative artificial intelligence for specific purposes, including identifying relevant law, recommending documents, filing supporting documents under certain conditions, and delivering final decisions. It mandates written consent from all parties for AI use and stipulates that errors resulting from AI use cannot be the sole basis to vacate, modify, or correct an arbitration award.
Proponents argue
Supporters argue that the bill modernizes the arbitration process by integrating advanced AI tools, which can significantly enhance efficiency, reduce costs, and streamline tasks like legal research and document management. They contend that requiring explicit consent ensures transparency and party autonomy, while the assumption of risk clause prevents frivolous challenges to awards, preserving arbitration's finality.
Opponents contend
Opponents might raise concerns about the reliability and potential for bias in AI, even non-generative systems, within legal contexts. They could argue that limiting the ability to challenge awards based on AI errors, even with consent, might undermine due process and fairness, particularly for parties less equipped to understand the complexities and risks associated with AI integration.
The bill is exceptionally short and clearly written, making it very accessible for a quick and thorough evaluation by any interested reader.
Section 2a(a) of Chapter 1 of title 9, United States Code
Authorization for AI Use in Arbitration
This provision amends federal arbitration law to explicitly allow parties to use artificial intelligence in arbitration for several defined purposes. These include identifying relevant laws and case law, making recommendations on document order and relevance, filing supporting documents when facts are undisputed and only a question of law exists, and delivering final decisions. It also includes a broad allowance for 'such other purpose as a party determines necessary'.
Supporters argue
Supporters argue this provision embraces technological advancement to make arbitration more efficient and accessible, allowing parties to leverage AI for tasks like legal research and document management, thereby potentially reducing legal costs and speeding up dispute resolution.
Critics contend
Critics might express concern that the broad 'such other purpose as a party determines necessary' clause could open the door to unforeseen and potentially problematic uses of AI beyond the enumerated specific functions, making it difficult to regulate future applications.
Tradeoffs
This provision balances the desire for technological efficiency and cost reduction in legal processes against concerns about the scope and potential misuse of AI, particularly with the open-ended 'other purpose' clause.
Section 2a(b) and 2a(c) of Chapter 1 of title 9, United States Code
Requirements for Consent, Disclosure, and Limitation on Appeals
This section mandates that artificial intelligence can only be used in an arbitration if all parties provide written consent and actively use such AI. Crucially, it also stipulates that a party who uses or agrees to the use of AI cannot later move to vacate, modify, or correct an arbitration award solely because an error occurred due to the AI's use.
Section 2a(d) of Chapter 1 of title 9, United States Code
Prohibition of Generative Artificial Intelligence
This provision explicitly prohibits the use of generative artificial intelligence in any arbitration conducted under this title. This means AI systems capable of creating new content, such as text or images, are not permitted, distinguishing them from analytical or organizational AI tools.