Contracting for AI: Mitigating Vendor Risk
A guide from Jackson Lewis details critical AI-specific contract terms to manage risk in technology procurement, including disclosure obligations, data use limitations, and liability.
This podcast transcript offers a guide for counsel on navigating the contractual risks of procuring AI technologies or services that use AI. Attorneys highlight the need to update standard agreements to address AI-specific issues, moving beyond traditional software contracts. Key negotiation points include forcing vendor disclosure of AI use, precisely defining "services" to cover AI-driven activities, and scrutinizing data-use clauses, particularly where vendors want to use client data for model training.
Sophisticated clients and counsel care because inadequate contracts can expose an organization to significant liability for data breaches, biased AI outputs, and regulatory penalties under frameworks like the CCPA or GDPR. The guide stresses that default vendor terms often disclaim liability for inaccuracies and limit indemnification to IP infringement, leaving the customer responsible for other critical risks. Counsel should push for broader indemnification, clear lines of responsibility for AI-generated deliverables, robust audit rights, and contractual requirements for human oversight in high-stakes decision-making contexts.