During the past several years, artificial intelligence has gradually entered the mainstream. It’s in phones, social media feeds, cars, apps, search engines, medical and smart home devices and other consumer electronics products. So, if AI is in all of these products, shouldn’t it be subject to the same types of protections that govern all products consumers use?
Of course, say responsible AI and technology advocates, but this has hardly been the prevailing opinion. Then, on April 19, the Federal Trade Commission issued a statement that throws down on the relationship between AI and consumer protection laws. “Aiming for truth, fairness, and equity in your company’s use of AI” sends an unambiguous signal that the FTC is now taking a hard look at the consumer protection implications of AI.
While it may seem difficult to draw a line between uses of AI and consumer protection laws, FTC makes it quite clear what they’re looking for, and where they’re looking. The first issue is bias and discriminatory outcomes. They say:
The question, then, is how can we harness the benefits of AI without inadvertently introducing bias or other unfair outcomes? Fortunately, while the sophisticated technology may be new, the FTC’s attention to automated decision making is not.
Issues such as unfair or deceptive business practices, fair credit reporting and equal credit opportunity are top of mind, supported by guidance on how FTC expects companies to comply with existing consumer protection laws. They focus first on data—particularly the imperative to design models that do not exclude marginalized, underrepresented or otherwise vulnerable groups. They caution businesses to watch out for discriminatory outcomes, whether in the context of advertising or allocating healthcare resources—like critical care or vaccine eligibility, for example. They warn against exaggerated claims—frequently a danger with new products or technologies, especially when they are as complex as AI can be. They urge businesses to do more good than harm.
But it’s the final caution that is most striking:
Hold yourself accountable – or be ready for the FTC to do it for you.
This may sound like a threat, and likely it is. Just because AI is a new ingredient in consumer products, that in no way excludes it from the same types of regulations that cover cars, blenders or your credit cards. The question is whether FTC’s bite will live up to its bark; we’ll just have to see about that.
All this being said, we in the United States are still living in an ad-hoc legal landscape for AI. This may not always be the case; as in the case of GDPR, the European Union is proposing the first-ever legal framework on AI. Will the EU take the lead from a global perspective? If GDPR was any indication, the answer is yes. And, if GDPR was any indication, the devil will be in the details.