Viral ‘joybait’ pranks use Meta’s AI glasses and facial recognition to ambush strangers

Meta's AI glasses have picked up a new line of work: helping creators stage fake feel-good encounters with strangers identified through facial recognition, all filmed hands-free. One such clip has passed 60 million views, and the kindness on screen is the least honest thing in it.

Meta AI Watch analysis

What happened

Joybait is ragebait's wholesome twin: content engineered to make you smile rather than seethe. Futurism's report traces the format through a viral video by the prank creator Levese, who spotted the aspiring novelist Travis Schick shopping with his partner at a Walmart. Rather than approach, Levese first photographed Schick's face on his phone, ran it through a facial recognition site and used the results to find his social accounts, which revealed him as an author. Only then came the approach, Meta glasses recording: a greeting by name and warm professions of love for Schick's novel The Redhead, delivered to a writer with no idea how his week had just been hijacked.

The reward structure is the story. The reel has drawn more than 60 million views and 30,000 comments, Schick's follower count jumped from around 150 to more than 12,700 within days, and well-wishers have been flooding the comments under his novel. Levese, with 225,000 followers, has posted dozens of videos in the same format, each pulling millions of views. The journalist James Vincent, who spotted the clip, called it 'individualized surveillance culture posing as feelgood content', arguing that whatever good reached Schick, the larger effect is a society that monetises kindness, exploits the goodwill of strangers and dismantles privacy in public. The glasses, only a few months on the market, had already been dubbed 'pervert glasses' for their ability to film people unnoticed; joybait gives that reputation a business model.

Why it matters

The scheme needs no hacking and no exotic kit: a phone camera, a public face search and a wearable that records without anyone noticing. That combination dissolves public anonymity at consumer prices, and the algorithm pays for it. The defence already appearing in the comments, that Schick gained readers and followers, is precisely the normalisation problem. Consent does not become retroactive because the outcome happened to be nice.

The practical lesson for anyone in public is uncomfortable. A stranger who knows your name has not recognised you; he has looked you up.

Our read

The smile is the product and the stranger is the raw material. Schick's windfall is real but incidental; he was cast without audition. Platforms now hold a clear policy question, because face-search ambush content sits awkwardly beside every rule on harassment and biometric identification. Treat viral kindness with the same scepticism as viral rage: if it is engineered to travel, someone chose the target.

What to watch

• Platform response
  Whether TikTok or Meta acts on clips that pair face searches with covert recording.
• The copycats
  Whether the format spreads further, given dozens of videos already sit in the series.
• The targets
  Whether Schick or future subjects respond, or explore legal routes over the filming.
• Meta's move
  Whether the company addresses glasses-enabled ambush content beyond the existing camera indicator.

Discussion spark: Schick ended the week with more readers and Levese with more views: does a good outcome for the target make the covert facial-recognition ambush acceptable, or is 'it worked out for him' exactly how surveillance becomes culture?

not affiliated with, endorsed by, or operated by Meta