How to Negotiate AI Clauses in Commercial Contracts

This guide for counsel outlines critical issues when negotiating AI-related provisions in vendor and customer agreements. As AI tools become ubiquitous, counterparties are inserting specific clauses to manage novel risks related to data privacy, intellectual property, system bias, and security. Standard terms on liability, indemnification, and compliance may not be sufficient, creating potential exposure for unprepared clients. The authors stress the importance of first understanding an organization's internal AI governance and capabilities before committing to contractual terms. Key negotiation points include restrictions on processing confidential or personal data, requirements for data not to be used in model training, and prohibitions on AI-driven decision-making. Counsel should also scrutinize attempts to create AI-specific carveouts from liability caps, broad 'compliance with all laws' warranties, and expansive audit rights. The guide advises carefully defining any obligations for human-in-the-loop review and verifying that both insurance coverage and subcontractor practices align with the contractual commitments being made.