It sounded at least somewhat promising when Tinder announced last year that it would start requiring photo verification for new users across the United States. Deploying a technology called “Face Check”, Tinder would have every such user take a video selfie whose coordinates would be compared against pictures uploaded with the profile. While this is a partial improvement over the previous state of things where facial verification was generally optional, significant flaws are emerging.
Earlier this month, journalist Christophe Haubursin uploaded a video to YouTube that showed a strange phenomenon whereby a lot of Tinder profiles exhibited a similar pattern without obvious initial explanation. The first eight profile pictures would show one individual, but then the ninth picture would be an artistic version of a different individual (say, a face embedded in a painting or mural). A reverse image search would uncover that the first eight pictures were of a–conventionally good-looking–person with a different identity than that portrayed by the profile.
Haubursin began suspecting that Face Check was deeming “verified” profiles where as few as one of the pictures corresponded to the required video selfie. He tested this hypothesis himself and confirmed it: all a profile needed to render it verified was one (even stylized) picture that matched the video selfie, even if it was preceded by eight pictures of someone who looked completely different and who most people would assume was the actual holder of the profile. Haubursin also contacted some of the people with these types of profiles, and unsurprisingly at this point they were actually crypto scammers.
I highlight this story for multiple reasons. One is to warn dating app users who encounter profiles of this type that the odds are greatly elevated of this being scammers. Another is to raise an issue I have highlighted in my previous work, such as most recently in my forthcoming article “Tinder Backgrounds“: when a dating app operator claims that a photo has been verified, we need to analyze the meaning of that statement to decide how much to trust the result of said verification. If Haubursin’s reporting is accurate (note also that he states that he has reached out to Tinder and not heard back), then Tinder’s current iteration of Face Check–by deeming approved a profile where eight of the nine pictures do not correspond to the profile owner–does not provide actual verification in the way that most people would use the term.
The situation is dangerous both for users’ safety and potentially Tinder’s risk of legal liability. Arguably, users could reasonably rely on Tinder’s claims regarding a verified profile and lose money or be harmed in some other way as a result. Cases such as Estate of Bride v. YOLO Technologies, Inc. (9th Cir. 2024) suggest that Section 230 may not shield online entities that promise safety measures that they do not truly take. Depending on how a court interprets what Tinder conveyed to its users regarding its photo verification process, this may legally not end well for the dating app.
As Haubursin states in his video, one seemingly straightforward solution would be for Tinder to mark individual pictures as verified rather than whole profiles. This would certainly shed clarity in the case of profiles where one picture is that of the scammer and the rest of an attractive individual meant to attract unwitting victims. With Congress increasingly taking notice of the significant role that dating apps play in the perpetuation of romance scams, dating apps will need to exercise much greater care if they want to avoid not only lawsuits but also future regulation.
