Google/YouTube have been criticized before for allowing deepfake celebrity scam ads to proliferate across their platforms, with fake Joe Rogans and Taylor Swifts telling people about life-changing supplements or free money from the government. After external pressure, the company took down many of these ads. What’s interesting about the “Pantry Find” ad is that the YouTube video itself does not contain a celebrity deepfake. In fact, Mormon Bob says outright that he can’t say “Bill’s” last name “because this video has already been taken down more than three times for that exact reason.” But when you click through to the site, there’s a new video, not hosted by YouTube, containing a realistically-faked Bill Gates telling you that the only way to preserve what’s left of your mind is to send the contents of your wallet to “MARKET AUTO EXPORT LLC” (or whoever is claiming to be them). YouTube may have minimized the use of AI deepfake celebrities on the platform, but it’s clear that there are loopholes.
YouTube has resisted calls for it to vet ads before they go up, rather than in response to user complaints. When the Liberal Democrats in the U.K. said the platform should have to conform to the same advertising standards as traditional broadcasters, the company said "YouTube is not a broadcaster and should not be regulated like one” promising that “we have strict policies that govern the ads on our platform which we enforce rigorously." Nevertheless, the U.K. is planning to require platforms to block scam ads, which will hopefully at least help in Britain.
I think it’s clear that vetting is not being done by hand. The Pantry Find would never have gotten “on the air” if someone had manually clicked through to the site, where they would have immediately seen it was a deepfake scam. I do not know how many YouTube ads are scams, or dubious products, but Help Net Security says YouTube is riddled with fraud, in both the paid ads and the ordinary videos. Sandeep Abraham, a business security consultant, confirmed to WBUR last year that “scam ads on social media and YouTube [have] really surged in the past couple of years.” Meta has had the same problem, with scam ads targeting seniors reportedly netting Meta millions of dollars in ad revenue, and being successfully reposted even after takedown.
Of course Google could stop this if they wanted to. They could have an employee review every ad before it’s posted to ensure it’s not a scam, or at least every ad containing certain keywords (Alzheimer’s, for instance). Variations on the celebrities-endorsing-honey-for-dementia scam have been circulating since last year across platforms, so it would be easy to spot if ads were pre-screened before being shown to people. But that would be time-consuming and costly. Google has a financial incentive to be lax about this stuff. Policing it is resource-intensive, and they get a cut of the scammers’ profits. Of course, ideally the federal government would step in, but since the president himself is a con man who has gutted consumer protections, this is very unlikely to happen. Hopefully crackdowns in Europe will result in some better practices filtering over into the U.S. Perhaps left-leaning younger workers at the big tech companies could exert internal pressure on their employers to reform. I’m sure many of them are uncomfortable with the fact that even a fraction of their salaries are paid by scamming dementia patients. They need to step up and demand a genuine crackdown on predatory advertising and bogus claims on the platform.
The Pantry Find scam makes my skin crawl. When I think about the type of person this is meant to go after, I am saddened and enraged. I don’t know who is responsible, but I know they should be prosecuted. I also know that the scammers themselves are only part of the problem. These schemes proliferate in part because of tech companies that don’t even exert the bare minimum responsibility of keeping elderly and cognitively impaired people safe from online predators.