As law firms and other businesses increasingly look to AI-driven software to drive efficiency, the importance of meticulous review of not just their capabilities and features, but also the agreements under which they are provided, becomes a crucial part of the onboarding process. Whether labeled as licenses, terms of service, or otherwise, these agreements define the risks associated with using AI services.

For example, in response to the legal industry's specific need for confidentiality of client information, providers are offering seemingly closed systems that claim to remove the provider's connection and access to the law firm's data with which the system works. But in some cases, the agreements may nonetheless enable provider access for regulatory purposes, potentially creating privilege issues.

In a broader context, companies aim to steer clear of claims of infringement arising from the system's use of training data or its generation of work product that mirrors existing copyrighted materials. While many agreements offer indemnification against such claims, these protections may come with significant limitations, or may only be available in circumstances where the risk of infringement would be minimal in any event.

As AI tools become increasingly integral for all businesses, understanding the legal frameworks that govern their use and associated liabilities is essential for mitigating risk when implementing AI technology.