Hey, AI hucksters. Dial down the hype. Lots of love, the FTC
Analysis The US Federal Trade Commission has re-upped its admonition to businesses that they should not exaggerate the role or capabilities of artificial intelligence in their products.
Artificial intelligence, or AI, explains FTC attorney Michael Atleson in an advisory, can be defined in a variety of ways, but the agency is concerned it has become a hot marketing term.
“[A]t the FTC, one thing we know about hot marketing terms is that some advertisers won’t be able to stop themselves from overusing and abusing them,” said Atleson.
One thing we know about hot marketing terms is that some advertisers won’t be able to stop themselves from overusing and abusing them
There is undoubtedly a lot of AI hype. The past year has revived hopes for intelligent chatbots – despite a repeat of the same sort of embarrassing output from ChatGPT and its progeny that derailed Microsoft’s Tay – and has shown that training models on artists’ work without permission can attract venture funding and lawsuits. We’re now at the point where there’s a supposed need to prepare for the coming of artificial general intelligence (AGI), or at least to pour funding into AI ventures pursuing hypothetical systems that have the cognitive abilities of humans without the disobedience and independent thought.
In response to OpenAI CEO Sam Altman’s suggestion that “a new version of moore’s law that could start soon: the amount of intelligence in the universe doubles every 18 months,” NYU professor and AI expert Gary Marcus said, “A new version of moore’s law that has arguably already started: the amount of hype around AI doubles every 18 months.”
Atleson says AI hype, seen in a variety of products including toys and chatbots, may represent exaggeration and deception. And if that’s the case, be warned … ominous pause … the FTC is watching.
“Marketers should know that – for FTC enforcement purposes – false or unsubstantiated claims about a product’s efficacy are our bread and butter,” said Atleson.
- Seeing as GPT-3 is great at faking info, you should lean into that, says Microsoft
- I can’t do that, Dave: AI drowns top sci-fi mag with story submissions
- Microsoft’s new AI BingBot berates users and can’t get its facts straight
- Take the blue pill: Keanu Reeves has had enough of AI baloney
This may not be quite the threat Atleson imagines given the agency’s reputation for toothlessness.
The FTC’s ineffectiveness in privacy enforcement is cited in a 2022 paper [PDF], “Fox in the Henhouse: The Delegation of Regulatory and Privacy Enforcement to Big Tech.” Authors William Bendix (Dakota State University), and Jon MacKay (University of Auckland), cite how Apple and Facebook between 2010 and 2020 were able to ignore FTC privacy enforcement for a decade because they have the technical capacity to conceal violations and the financial resources to absorb fines.
“Initially, both firms decided to ignore the agency’s enforcement requests and let app developers commit ongoing legal and privacy violations,” the paper explains.
“Apple allowed developers to trick children into making unauthorized in-app purchases on their parents’ accounts. Facebook, meanwhile, gave developers broad access to its users’ personal data, even after assuring users their information was protected. Eventually Apple complied with FTC orders in full, while Facebook committed ongoing policing failures.”
Trump-appointed now-former FTC Commissioner Christine S. Wilson, who resigned this month in protest of regulator boss Lina Khan’s “disregard for the rule of law and due process,” noted the agency’s perceived ineffectiveness in a January 26, 2022, speech [PDF] assessing the Biden administration’s approach to antitrust enforcement: “On the consumer protection front, critics characterized thirty years of Made in USA enforcement as toothless because consent decrees with injunctive relief – but not monetary relief – were common.”
And – consistent with her Republican party’s preference for low-impact regulation – she rejected the notion that the FTC is all bark and no bite, noting that recidivism is rare and proving harm is difficult.
Khan do attitude
The FTC under Khan has tried to be more assertive by blocking allegedly anti-competitive acquisitions, such as Meta’s pursuit of Within Unlimited. But Meta prevailed in court and it’s not clear the agency’s other high profile antitrust case, to block Microsoft’s absorption of Activision Blizzard, will succeed. As Khan herself suggested in a 2017 law review article, today’s antitrust law doesn’t really provide the tools to bring Big Tech to heel.
The agency may find that keeping marketers honest under its consumer protection mandate is easier than antitrust enforcement, though critics have argued the FTC has failed to bring social media influences to heel [PDF]. The regulator subsequently revised its advertiser rules to better address the problem. Yet the fake reviews and undisclosed social media shilling continue.
The federal trade watchdog has been steadily trying to tame AI excess. In 2020, it issued AI guidance telling companies to avoid algorithmic discrimination, to avoid deceiving folks about the use of automated tools, and to be transparent about the collection of sensitive data.
The following year, the agency made clear it would go after firms using biased algorithms under the deceptive practices portion of the FTC Act, the Fair Credit Reporting Act, and the Equal Credit Opportunity Act. Last year, the agency issued a report to Congress calling for the development of a legal framework to ensure AI tools do no harm. And this year, the FTC has reminded marketers to avoid unfounded claims about AI.
The FTC no doubt will be busy. Check back in a year or two to see if it’s been effective. ®