Amid the recent wave of artificial intelligence (“AI“) hype in the context of various products, the Federal Trade Commission (“FTC“) has issued a guidance warning companies against overusing and abusing the AI term in marketing their products. In its guidance, the FTC notes that AI is an ambiguous term with many possible definitions, and warns that false or unsubstantiated claims about a product’s efficacy are likely to attract its attention. In particular, the FTC highlights the following guidelines and warnings with respect to the marketing of AI products:
- Companies should not overpromise what their algorithm or AI-based tool can deliver. Performance claims that lack scientific support or apply only to specific types of users or conditions may be considered as deceptive by the FTC. Similarly, companies should also refrain from making comparative claims, promising that their AI products are better than non-AI products, without adequate proof for such claimed advantage;
- Companies are advised to ensure that they are aware of the reasonably foreseeable risks and impact of their AI products before marketing them, noting that in case their products fail or yield biased results, they will not be able to blame third-party developers or say they are not responsible because the technology is a “black box” they do not understand;
- Companies are strongly warned against making false claims about a product being AI-enabled. The FTC may investigate and examine the product to verify whether it is actually using AI technology. It is being emphasized by the FTC that using AI tools in the development process does not necessarily make the product AI-powered.
This guidance joins a previous guidance published by FTC in April 2021, which provided principles for reducing bias and unfair outcomes in AI, particularly with regards to discrimination against legally protected classes such as race, gender, and religion. These principles included ensuring a solid data set that accounts for data gaps and limitations, testing algorithms to prevent discriminatory outcomes, embracing transparency and independence, inter alia, by using transparency frameworks and independent standards and conducting independent audits and inspection, being careful about how data is acquired and used, and generally ensuring that the AI model does more good than harm.
According to the FTC, failure to comply with the above principles and guidelines regarding the responsible use and marketing of AI products, may result in enforcement based on several federal laws including the FTC Act, the Fair Credit Reporting Act and the Equal Credit Opportunity Act.
Author: Ariel Yosefi,Oded Kramer.