On the internet, how do you prove that you are real? From dating sites to co-opted communities, verification techniques are multiplying to hunt down fake profiles, in order to stem the excesses of online anonymity, between video selfies and sponsorships.
“On the internet, no one knows that you are a dog”: the warning from 1993 in a cartoon in the cultural magazine The New Yorker still resonates 30 years later in France in the Assembly, where parliamentarians are examining the draft law on “digital security”.
Among the avenues studied, the idea – approved by the French Minister Delegate in charge of Digital, Jean-Noël Barrot – that social networks can offer their users to submit proof of identity, so that they can choose not to communicate only with other certified users.
In recent years, major dating sites have already implemented photo verification tools on a voluntary basis. Users can limit their searches to verified profiles only, which incentivizes adoption of these tools.
Bumble, one of the most widely used dating apps in the world, asks the user for a selfie imitating one of a hundred randomly provided poses, the company says. This selfie will be compared to the photos he posted.
The French Happn, also in the Top 10 worldwide, has been developing the certification of profile photos by video for two years.
“If you post a photo, you can make a video and the device, validated by the CNIL (the institution guaranteeing the privacy of the French, editor’s note) allows biometrics to verify that you are a real person,” explains Karima Ben Abdelmalek, president of Happn, to AFP.
A system also implemented since April by Tinder, one of the brands of the global sector giant Match Group, which also allows users to choose to only see certified profiles and even ask their “match” to verify their profile before discussing.
– Cooptation and rating –
The group is even considering launching verification by identity document, but without an announced date.
“Across Match Group (Tinder, OkCupid, Meetic, Match, Hinge…), 44 spam accounts (which do not respect the platform’s rules, editor’s note) are deleted on average every minute across the “all brands, and almost 5 million bots (automatic accounts) and spam accounts were deleted between January and March of this year, before having access to the platform or shortly after registration”, specifies Match Group to the AFP.
Happn also uses artificial intelligence to spot inappropriate behavior and automated spam. “If someone reacts a hundred times on a profile, while on average a human reacts 30 times, then he is not a human: he is a robot,” explains Karima Ben Abdelmalek.
The tool also detects those who ask for money or publish a phone number, and bans them in a few seconds.
Rather than sophisticated verification techniques, other sites use more traditional ways to avoid fake profiles.
A sponsorship site like Gens de Confidence (location) relies on co-optation: you have to collect several sponsorships from members to be accepted. The idea is that they know each other in real life, the ultimate guarantee of trust.
Sponsors are also held responsible subsequently: they can have their account deleted in the event of bad behavior by a godson during a rental. Over the years, the site has expanded its audience but the sponsor system is supposed to maintain a guarantee of the reliability of the profiles.
The same applies to all sites where users rate each other, such as BlablaCar, or even rate businesses and services. Without certifying that these are real profiles, this system creates a co-opted reputation, supposed to eliminate usurpers.
“We are moving towards a system where certified and non-certified profiles will coexist, like on AFP.