Bridget McCormack, president and CEO of the American Arbitration Association-International Centre for Dispute Resolution (AAA-ICDR), is integrating artificial intelligence into the organization's core operations and dispute resolution services. The move introduces enterprise-grade AI tools and a specialized AI arbitrator for construction disputes, setting a precedent for how legal institutions can adopt automation while maintaining strict security and accountability standards.
Enterprise AI adoption and internal training
The AAA operates an enterprise ChatGPT account to maintain a secure environment for processing internal documents. McCormack said the organization requires all staff to complete generative AI training and sets specific AI usage goals. "Everybody has generative AI goals. We have guidelines for our staff to follow. There are things they're allowed to do and things they are not allowed to do," she said. This internal foundation supports external tools, including a multilingual client chatbot trained on the organization's historical rulings, rules, and cases to assist parties who cannot afford legal representation.
The AI Arbitrator for construction disputes
The organization's most significant development is the AI Arbitrator, a tool designed to accelerate claim parsing, document summarization, and analysis. Currently, the tool is restricted to two-party, document-only construction cases due to the high volume of disputes and a deep catalog of historical awards for training. "In ongoing construction projects, one dispute can stop the project, so they really need some mechanism for the disputes that come up along the way, to get those decided - and quickly too," McCormack said. A human arbitrator reviews the output and issues the final award to maintain accountability.
Governance, security, and public audits
To build trust, the AAA built the AI Arbitrator on-premises rather than in the cloud, and the system does not retain user data. The organization established a governance committee comprising lawyers, in-house counsel, and the chief information officer to evaluate internal builds and external AI partnerships. Furthermore, the AAA will conduct organized audits of the tool's performance and potential bias, making the results public to demonstrate transparency. "I think trust is what you hopefully earn by building transparently," McCormack said. "So we're showing everybody what we're doing."
Why this matters for legal professionals
Legal professionals evaluating automation should note the AAA's strict operational boundaries. By restricting the AI Arbitrator to document-only cases, requiring human sign-off on final awards, and publishing bias audits, the organization provides a clear blueprint for AI for Legal implementation. This approach balances efficiency gains in dispute resolution with the accountability and security standards that clients demand.
Your membership also unlocks:
