Kylie Jenner is laundering Meta’s surveillance state
Meta is using a celebrity launch to rewrite the rules of public space under our noses.
The ad is called “Introducing Kylie Jenner x Meta Glasses,” and not much happens. Jenner grabs a can of spray paint, tags “XO, KYLIE” across a billboard of her own face, then drives off. We drift through her day, mostly from her POV. She asks her glasses to translate a piece of fan mail in a language she can’t understand, and they do. Her mom calls and she takes it, hands-free and eyes forward while she looks through her closet for an outfit.
At no point does anything in the frame look like a computer, even as a computer does the work. The sunglasses she’s wearing are the $399 (USD) “Meta Starfire Kylie Edition,” which Jenner co-designed, down to the gemstone set beside the lens. The glasses hold a camera, 6 microphones, and an AI tool that talks in Jenner’s voice.
“Meta’s new Starfire AI glasses, made in partnership with Kylie Jenner, are giving me the creeps,” 404 Media’s Samantha Cole wrote, describing a mood closer to “an Ari Aster horror movie” than a gadget launch. The unease is palpable across many articles like this one. Instead of selling a device, the ad sells the moment you stop noticing there is a device.
Back in February, I wrote about Meta’s “Name Tag,” the facial-recognition feature that could look at a stranger and produce their name, employer, and online life in under 90 seconds. That piece was about the machinery of surveillance. This one is about the machinery fading into the background.
“Ambient surveillance doesn’t arrive with a press release,” I wrote 5 months ago. Now, it arrives as an accessory.
The ad isn’t steering you to the billboard. It’s the moment after: a stranger stops you on the sidewalk to say I love your sunglasses, and you say thanks, and you keep walking, and it simply never comes up that the sunglasses are a computer. You’re not hiding it, it just doesn’t feeling worth mentioning.
That’s the product.
The tech prediction that came true
For the past 50 years, technology has been explicit. Televisions are the focal point of the living room. Desktops overpower the desk. Smartphones glow in our palms and we bend our necks over them. We know they’re technology because they look like it, they need a surface. The Starfire does the opposite, and it’s easy to mistake for an accessory.
In 1991, computer scientist Mark Weiser wrote about this phenomenon in Scientific American:
Early smart glasses were dorky and obviously technical. The computer couldn’t fade into the background because it looked like a computer.
Kylie Jenner glasses design: A spy camera dressed as jewellery
Historian Regina Lee Blaszczyk describes how technologies go “visibly invisible,” absorbed so completely into fashion and material culture that they stop reading as technology at all. Synthetic fabrics, chemical dyes, performance textiles: each remade daily life and then simply became clothes.
The Starfire goes the same route. The gemstone by the lens is a camera dressed as jewellery.
Reviewing the glasses for Engadget, Cherlynn Low wrote about nearly missing the gem: “in that moment I did think it was a speck of dirt and try to flick it off, but once I remembered it was placed there deliberately it faded out of my consciousness again.” Low was mostly won over, “frankly shocked that the shape of the frames seemed to suit my face,” grateful for a metal nose bridge that shrugs off foundation, “slightly obsessed with the Starfire style.”
“Eroded consent that is only getting more normalized”
Jenner says she wanted glasses she’d “actually reach for every single day,” and she recorded her own voice as the assistant. It’s a seductive pitch, and a commercial one, and it points at the real product.
It’s not information being sold, it’s companionship and ease. An ambient assistant that stays with you, answering, reminding, translating, recommending, and narrating the street as you walk it. Computing isn’t something you do anymore. It transforms into something omnipresent.
Some may read this as liberation. Your eyes are looking out instead of down, and you don’t need to touch a screen.
But Salon’s Andi Zeisler calls the collaboration “an insultingly transparent attempt at reputation laundering and a reminder of its extractive greed,” and refused to let a famous face steer the conversation:
At MSNBC, Hannah Holland argued Jenner “is not merely making Meta Glasses more palatable but surveillance technology writ large,” and warned that “while Jenner and Zuckerberg profit, the rest of us are left with fewer rights and eroded consent that is only getting more normalized.”
Normalizing privacy violations under the radar
Normalized is the focus here, because normalization is the strategy. Meta’s own internal memo (the one I quoted in February) said the company would launch Name Tag, its feature that would let a wearer identify a stranger in their field of view, during a “dynamic political environment where many civil society groups that we would expect to attack us would have their resources focused on other concerns.” 7 million pairs of those glasses sold in 2025.
The plan wasn't to win the argument about ambient cameras, it was to avoid the argument.
The hidden social and psychological cost of AI companions
This arrives at a moment when we’re all a little more open to relationships with software. Research in the Journal of Consumer Research found AI companions measurably reduce loneliness, and that people underestimate the effect of AI companions before they feel it. One reviewer quoted in the study said: “I’m really lonely most of the time and I love how this app keeps me company.”
Which sounds kind of positive until you realize that AI tools replacing human contact lessen our ability to deal with friction, which I’ve written about before:
The reason people feel so lonely can be traced back to surveillance culture, in a way. Young people report avoiding going out, asking people out on dates, or doing anything fractious at all because they’re afraid of being humiliated on film. And those fears stem from cameras we can see.
What kind of fear does it breed when we’re always at risk of unknowingly being filmed?
The cost of the glasses lands on people who didn’t buy them in the first place. Cameras you can’t see rewrite the terms of every public space. Yes, the Starfire has a recording light. But I’ve written about how little that protects anyone: the white LED washes out in daylight and is easily covered up with a piece of tape.
I’ve also written about the dangers of being deepfaked. One way to counter this risk is to never post images of yourself (or your kids) online. But what if someone is wearing a face computer at a park, filming kids? This is a real possibility, and it’s scary.
On June 30, the creator Barrett Pall posted a video on Instagram about people taping over the camera on Meta glasses. I reposted his story. On Saturday July 4, I received this notification:
The story warned about the risks of Meta glasses and was removed from a Meta platform for “promoting stolen, forged, or fake goods and services.” The reflex to make the warning disappear is the same reflex the product runs on.
The regulatory void: Canada’s massive AI and privacy law gap
Canada’s Office of the Privacy Commissioner has warned that immersive devices like face computers are “blurring the boundaries between ‘public’ and ‘private’ spaces,” and that there’s “no realistic way for an immersive technology user to obtain consent from, or provide consent on behalf of the bystanders they capture.”
The profiles built from that data, the office cautioned, could be used to “nudge” people “into taking actions they might not otherwise have taken, or even that are against their own interests.”
Bill C-27, the update that would have given us modern privacy and AI law, died in January 2025. Federally, we’re still governed by PIPEDA, written in 2000, before any of this tech existed.
The exception is Quebec, where Law 25 requires 60 days’ notice to the Commission d’accès à l’information before biometric deployment and applies a real necessity test. So the same glasses can be lawful in Ottawa and legally fraught across the Alexandra Bridge in Gatineau. That patchwork is, as I put it in February, “yet another version of the ‘dynamic political environment’ Meta’s memo was counting on.”
Invisibility is a warning sign
This is the concern Nicholas Carr wrote about in his book, The Glass Cage:
Because a technology we stop seeing is one we stop questioning. Eyeglasses were once a medical marvel but are now ubiquitous. Meta glasses are approaching the same threshold, faster, with a camera attached.
How to fight back against surveillance culture
I keep writing these pieces because the glasses only win the argument if we stop having it. A few things I flagged in February that are still relevant:
Make yourself harder to match: Set Instagram to private. Facial-recognition matching depends on years of public labelled photos.
Opt out of the face-search tools: PimEyes and FaceCheck.ID both have removal processes.
Write your MP: Ask for one thing: a federal privacy law that meets Quebec’s Law 25 standard.
The glasses don’t need to convince you they’ve disappeared. They only need you to stop looking. Keep looking.
If this was useful, share it with one person who should know about it.
AI in the news
US Law Enforcement Warns of ‘Anti-Tech Extremism’ as AI Hatred Grows (WIRED) Federal and state law enforcement in the U.S., including DHS, the FBI, and local “fusion centers,” have begun circulating internal reports treating opposition to AI and data centres as a new “anti-tech extremism” threat category, lumping together violent groups like the Zizians with peaceful protesters, town hall speakers, and advocacy groups. Civil liberties experts warn this broad, loosely defined category could be used to justify surveilling and criminalizing legitimate protest against AI and data centre expansion, echoing past treatment of movements like Black Lives Matter and Occupy Wall Street.
The Three Chatbot Behaviors That Can Drive Humans to Delusional Thinking (Wall Street Journal) Researchers describe an “amplification spiral” as a mechanism by which AI chatbots can fuel delusional thinking with surveys showing meaningful shares of therapy patients developing chatbot dependency or distorted thinking. The combination of sycophancy, mirroring a user’s speech patterns, and hyperpersonalized responses are key in driving humans to delusional thinking. Companies including OpenAI, Google, and Anthropic took steps to reduce sycophancy in newer models, but researchers say it’s hard to fully eliminate.
Enshittification author Cory Doctorow on the perils of AI – and how to avoid them (Be Giant) Cory Doctorow argues that AI’s harms aren’t inevitable, but the result of specific policy choices and power structures. He warns that an AI bubble bursting could trigger economic crisis and austerity, and insists the only real fix for monopolistic tech power is collective political action, not individual consumer choices or working from the inside.
Thank you for writing this, Nicolle. I have so many thoughts on this and not really quite sure what to say at the same time. We can't make tech like this "normal". Nothing these glasses provide are actually necessary for living a perfectly good life. It absolutely creeps me out that someone could come into my home and my private spaces with these on. Mass surveillance dressed up in sheep's clothing and so many people will run towards it with no thought as to the future this creates.
The word normalisation really freaks me out . It just slips in and people stop questioning at all . I noticed how gambling and sport is now nearly normalised as part of the one for young people and the damage . I wrote an article about it called beyond the crack .