On January 9, 2024, the Federal Trade Commission published a blog post reminding artificial intelligence (“AI”) “model-as-a-service” companies to uphold the privacy commitments they make to customers, including promises made in Terms of Service agreements, promotional materials and online marketplaces.

“Model-as-a-service” companies develop AI products and make them available to customers via end-user interfaces or application programming interfaces (“APIs”). An example of a “model-as-a-service” company the FTC describes is a business that develops large language models (“LLMs”) and sells access to them to online retailers for use as customer service chatbots.

The FTC notes that AI “model-as-a-service” companies have a business incentive to continuously improve their products by training AI models on new data sets. The FTC cautions that if companies choose to retain consumer data for AI training purposes without clear and conspicuous notice or obtaining affirmative express consent, they are at risk of violating consumer protection laws. The post reminds companies that both misrepresentations made in privacy commitments and failure to disclose material facts to consumers are subject to FTC enforcement.

The FTC further notes that AI companies that misuse consumer information to undermine fair competition based on false privacy promises to customers also run the risk of violating federal antitrust laws.

While currently, there are no AI-specific federal regulations for the private sector, the FTC reminds AI companies that they must comply with existing federal laws and must uphold privacy commitments made to their customers.