AI In Service Provider Agreements
Artificial intelligence (“AI”) has moved quickly from a specialized capability to an everyday operational tool used across industries. Even service providers that do not offer AI-based products or AI-driven services frequently rely on AI behind the scenes—using tools such as chatbots, drafting assistants, analytics engines, or automation platforms to support their work. As these tools become embedded in ordinary business processes, they introduce risks that traditional commercial contracts were not designed to address. To manage these risks, companies are increasingly incorporating AI-specific representations and warranties into service agreements, ensuring that any AI used incidentally in service delivery is properly disclosed, authorized, and governed.
Customers entering into agreements with service providers should consider including certain representations in their agreements:
(1) Requiring Disclosure and Authorization of AI Use. A typical provision may require the service provider to affirm that it will not use AI technology in performing the services unless expressly disclosed in the agreement. If AI tools are permitted, the provider may be required to confirm that it has all necessary rights, licenses, and consents; that it has complied with applicable laws and industry standards; and that it has met all obligations under the terms governing any third-party AI tools it uses.
(2) Ensuring Lawful, Ethical, and Compliant Use of AI Tools. Service providers may be asked to warrant that any AI technologies they use—if allowed—were lawfully sourced, trained, or implemented, assurance that only permissioned data was used for training, and that such use does not infringe third-party rights or expose the customer to regulatory risk. Providers may also need to confirm that they maintain responsible-use practices, including internal policies for oversight, bias mitigation, ethical use of AI systems and for confirmation that employees follow established data-use policies. Where the customer maintains an AI or vendor-use policy, the agreement may require the provider to comply with those standards as a condition of service performance.
By incorporating these types of representations and warranties even in agreements where the provider is not delivering AI-based offerings, companies can reduce the risk that undisclosed or improperly managed AI tools enter their operations. This approach helps ensure visibility into the use of AI by vendors, provides contractual recourse if issues arise, and supports broader governance efforts as organizations navigate the rapidly evolving legal and regulatory landscape surrounding AI.
The information contained in this article is not legal advice and has not been updated to include changes in the law since this article was originally published.