The Growing AI Conundrum: Fake Reviews Edition

Online reviews have never been more important to consumers. According to some reports, the overwhelming majority of consumers not only read online reviews, they trust them to inform purchasing decisions. Given this powerful impact on consumer behavior, it’s perhaps no surprise that dishonest businesses are taking advantage of this tool, leading to an explosion of fake reviews on the internet. But unlike some other forms of dishonest marketing, it can be almost impossible to weed out the fake reviews from the genuine ones, particularly as AI technology enhances the ability of disreputable businesses to create persuasive, fabricated reviews at the snap of a finger.

This is undoubtedly a massive problem, not only for consumers and honest businesses, but for regulators as well. In an effort to rein it in, the FTC has taken some recent steps in the fake reviews sphere, with one being less contentious than the other. In August, in a 5-0 commissioner vote, the agency finalized a fake reviews rule that, among other things, prohibits fake reviews, including those that are AI-generated (the rule just took effect in October). Then in September, in a divided 3-2 commissioner vote, the FTC brought an enforcement action against a company called Rytr regarding its AI-enabled “writing assistant” that generates genuine-sounding yet wholly fabricated reviews of products and services.

The FTC’s complaint alleges that Rytr’s AI-generated reviews tool can turn elemental user input – just a product name – into “detailed reviews that contain specific, often material details that have no relation to the user’s input.” Based on these facts, the FTC took the position that Rytr violated section 5 of the FTC Act (the FTC’s new fake reviews rule was not yet in effect) – not because of a theoretical possibility that its reviews tool could be used to deceive consumers, but because fabricating reviews that sound like authentic experiences of genuine, bona fide users is the very reason for which it is being bought and sold.

It is important to note here that there is simply no economic incentive for consumers to buy a reviews tool to write their own honest reviews. Which may explain why Rytr’s marketing doesn’t appear to target consumers at all, but rather focuses on companies, marketers and the like.

The FTC’s action resulted in a proposed consent order that prohibits Rytr from advertising or selling its AI-generated reviews tool, one of its more than 40 AI-enabled “writing assistant” services, which the company calls “use cases.” The rest of Rytr’s services, including the company’s numerous other marketing tools, are unaffected by the proposed consent order, which the commission must now decide whether or not to finalize.

Both Commissioners Holyoak and Ferguson voted against the Rytr action and issued dissenting statements. Among other things, the dissenting statements expressed concern that a “possibility” that Rytr’s tools could be used for fraud is not in and of itself a violation of section 5 (“A consumer could use it to draft an honest and accurate review”), and that the FTC lacked evidence of consumer harm (“the complaint makes no allegation that any drafts were ever posted online.” Emphasis in original.)

Certainly, these points are worthy of consideration – an exploration of all views and analyses can only be beneficial and lead to a stronger end result, particularly where, as here, the case presents significantly complex issues. But they are not reasons to run. As we stated in a public comment filed with the FTC:

this case brings to light a massive, Daedalian problem that requires the implementation of practical and effective measures that are sensitive to the complexities of the deceptive practices at issue while also addressing the significant damages such practices can inflict if left unchecked. It is simply not sufficient to take a caveat emptor approach in the face of such guileful practices. Strong, coordinated regulatory action is paramount if the FTC is to fulfill its mission of preventing business practices that are deceptive or unfair to consumers and honest businesses alike.

In other words, this is not the time for regulators to simply agree to disagree, fold their hands, and move on. Rather, this is a time when consumers and businesses need the agency the most – to tackle a burgeoning and harmful issue before the monster is fully and irretrievably unleashed.

For this reason, TINA.org filed a comment today in support of the FTC’s efforts in protecting consumers and honest businesses from the proliferation of false and deceptive reviews used in marketing, and urging the commission to tackle this serious and far-reaching issue with vigor and solidarity.

UPDATE 12/18/24: The FTC approved the proposed consent order in the Rytr case and entered a final order.