Source: Holland & Knight

Holland & Knight provides a regulatory assessment of the growing practice of offering performance guarantees for Artificial Intelligence systems. The article highlights that as AI becomes more integrated into critical business operations, customers are increasingly demanding assurances about its reliability, accuracy, and effectiveness. In response, AI vendors are beginning to offer performance guarantees, service level agreements (SLAs), and warranties. However, these promises are attracting the attention of regulators, such as the Federal Trade Commission (FTC), who are focused on preventing deceptive or unfair marketing practices.

The publication advises that companies making AI performance claims must ensure they have a reasonable basis for these guarantees, supported by robust and objective testing data. Vague or unsubstantiated claims about AI capabilities can expose a company to significant legal and reputational risks, including regulatory enforcement actions and false advertising lawsuits. The article emphasizes the importance of clear and precise language in contracts and marketing materials, accurately defining the scope of the guarantee, the metrics used for measurement, and the remedies available if the AI system fails to meet the promised performance levels. It serves as a guide for AI developers and vendors on how to structure their performance guarantees in a way that is both commercially attractive and legally compliant.