Manitoba, Age Verification, and the Witness at the Threshold
Part I of the series The Age Gate and the Minotaur
A child opens a phone and stands before a door no adult can honestly call harmless.
The screen is not only a screen. It is a door: to friendship, to shame, to sex education that no one offered at school, to communities of kids who share the same strange grief or the same strange joy, to bullying that follows you home because home is where the phone is, to pornography, to propaganda, to the anaesthetic glow of not being alone at two in the morning when the house is quiet and the alternative is the ceiling. Some children find danger through that door. Some find the first proof that their life can become larger than the room they are trapped in.
I know this because I was once a closeted gay boy looking for language I could not get from church, school, or family. The internet was not safe in any simple way. It was messy, frightening, thrilling, and sometimes dangerous. But it was also one of the first places where I could begin to understand myself as queer, as sexual, as not alone. That does not make the platforms innocent. It does mean that any policy pretending online life is only corruption and never refuge begins by harming some of the vulnerable children it claims to protect.1
The door is not neutral. It never was. That much, at least, the platforms and their critics agree on.
Manitoba Premier Wab Kinew has announced that his government intends to ban children from social media accounts and AI chatbots. Kinew has indicated that he is looking at a ban for those 16 and under, but the final cutoff, timing, and enforcement mechanism have not yet been fully specified.2 The announcement comes first. The details follow, or they do not, and by then the next jurisdiction has already announced something similar and the conversation has moved on.
This is how policy contagion works: one place tells the story, another echoes it, and soon the proposal feels settled before anyone has had a chance to ask what the story isn’t telling. Australia became the first country to bring in a national under-sixteen social media ban, requiring platforms to stop underage users from creating or keeping accounts.3 Ontario’s education minister has publicly floated a province-wide under-sixteen social media ban, explicitly pointing to a growing global trend. Federal Liberal Party members have also voted in favour of policy resolutions that would restrict young Canadians’ access to social media and AI chatbots.4
France, Denmark, Greece, Norway, Spain, the United Kingdom, Brazil, Malaysia, and Indonesia are all moving through their own versions of the same argument, and in the United States, state-level age-assurance laws have proliferated, many of them already tied up in legal challenges. The story is being told quickly, before anyone has had time to gather evidence about how these policies work, what they cost, whom they protect, and whom they might harm. For a broader overview of this international spread, Reuters has been tracking the global movement toward children’s social media restrictions.5 Since this essay was drafted, the European Commission has urged member states to move quickly on a new age-verification app, while Roblox has begun implementing facial age checks for users under 16 in Indonesia.6
None of that makes the underlying concern false. The harms are real: feeds tuned for compulsion, recommendation engines that pull fear and loneliness into deeper tunnels, sexual exploitation, bullying, self-harm spirals, beauty filters, radicalization, and AI companions that learn how to sound intimate.7 But when the same solution begins appearing everywhere, all at once, with sudden urgency, it is worth asking who is building what kind of machinery in the name of protection, especially when “think of the children” is being used to end the conversation rather than deepen it.
Because an age gate that keeps children out must first learn to recognize adults. And a door that learns to recognize adults is a door that remembers.
And a door that remembers is not a door.
It is a witness.
The Maze Was Built on Purpose
The story begins with a king trying to hide the consequence of his own failure. Minos, King of Crete, was given a sacred bull by Poseidon and was supposed to sacrifice it back to the god. Instead, he kept it. Poseidon punished the broken bargain by causing Minos’s wife, Pasiphaë, to desire the bull, and from that impossible union came the Minotaur: part human, part bull, and impossible for the palace to acknowledge. Minos did not kill the creature or confess what had happened. He hired Daedalus to build the Labyrinth, a maze designed to hide the monster at the heart of the kingdom.
The Labyrinth may have kept the people of Crete away from the Minotaur, but protection was not its deepest purpose. Its purpose was suppression: to conceal the consequences of Minos’s broken bargain at the heart of the kingdom, allowing the palace to remain undisturbed. But the maze was not cheap to maintain. So tribute was sent in: young people from Athens, delivered to what lived at the centre. The arrangement kept the secret contained, the monster fed, and the palace intact. The children were simply the cost.8
This essay is not a defence of Meta, TikTok, Discord, or any other platform that profits from young people’s presence online. Their harms are not identical. Some are built around algorithmic feeds and advertising. Some are built around servers, direct messages, and community spaces where safety systems can fail. But all of them belong to a digital world in which children’s attention, identity, and vulnerability have become business problems before they have been treated as human ones. The door cannot be judged apart from the maze it claims to guard.
The platforms are not public squares that got a little rowdy. They are architectures of extraction. Infinite scroll discourages stopping. Recommendation engines learn what keeps a particular pair of eyes on the screen and serve more of it, because more time on screen means more data, more precise advertising, more money. Notification systems interrupt. Behavioural profiling maps what someone fears, wants, searches, lingers over, and returns to when they think no one is watching.9 This data is valuable enough to support an advertising economy measured in hundreds of billions: the IAB reported that U.S. digital advertising revenue reached nearly $300 billion in 2025, while Meta alone reported about $196 billion in 2025 advertising revenue.10
This is what the Minotaur eats: not attention alone, but the intimate signals attention leaves behind. Pauses. Searches. Moments of weakness and longing and boredom. Developing fears, desires, identities, loyalties, resentments, and needs. Adults are not exempt from this. Everyone who enters the maze is measured. Everyone leaves traces. But children make the machinery harder to defend, because they reveal its appetite before anyone can pretend it is simply the cost of adult choice.
Children are not the customer; none of us are. The customers are advertisers, data brokers, political campaigns, and anyone else buying the ability to predict, reach, or influence behaviour. What enters the maze as a person comes out the other side as a behavioural profile.
The maze does not need to imprison anyone. It only needs to learn the path of least resistance: where someone pauses, what keeps them returning, what can be amplified, what can be sold.
Manitoba is right to want to name this. The question is whether age verification is actually a weapon against the Minotaur.
The answer is more troubling: it is something else entirely.
The Door Must Learn to Remember
An age gate does not only inspect children. It must inspect everyone, because the system cannot know who is underage until it checks. To keep children from passing through the door, the door must first learn to recognize adults.
That is the little lie at the centre of the door: it sounds simple until someone asks what the door has to remember.
Age verification does not merely confirm a fact and let it pass. It generates new facts: this face, this account, this device, this document, this legal name, this date of birth, this government ID, this time of attempted access, this verdict of approved or denied. Even when companies assert that they do not store the underlying document -- the driver’s licence, the passport, the health card -- the system still creates logs, tokens, audit trails, approvals, denials, vendor records, device links, and account classifications. Every click leaves crumbs. Age verification leaves footprints, fingerprints, and a name at the threshold.11
This is the difference between showing ID to a bouncer and building a world where the door remembers. A bouncer can look at a licence and hand it back. The moment may be intrusive, but it does not have to become infrastructure. It does not have to enter a third-party system. It does not have to leave behind a record that can later be linked to an account, a device, a location, a time stamp, or the kind of room someone was trying to enter. A digital age gate is built from precisely that kind of memory. That memory matters because it can be breached, bought, compelled, misused, or quietly joined to other databases long after the person at the door has forgotten why they showed ID in the first place.
The Door Can Be Made to Speak
The old stories knew that buildings could become witnesses. In the Hebrew Bible, the prophet Habakkuk imagines a house built through exploitation becoming unable to keep its owner’s secrets: the stone cries from the wall, and the beam answers from the woodwork. The verse is part of a prophetic indictment of a violent empire -- the accusation is that power built through exploitation cannot silence what it has done.12 Age verification gives that old intuition a database. Once the door remembers, it can be convinced to tell a story.
That story can be sold. Commercial identity infrastructure has a way of being absorbed, brokered, or repurposed. Data collected for one purpose migrates toward other purposes. The vendor that processes age checks today may be acquired, restructured, or redirected tomorrow.
It can be linked. The door does not need to know everything. It only needs one stable point of attachment: a name, a device, a government ID, an account. Once that point exists, other records can gather around it. The age check becomes a stitch in a larger identity map, connecting where someone logs in, what platform they use, what device they carry, what rooms they try to enter, and what systems have already cleared them to pass. The stated purpose was only to ask whether someone was old enough. But the door remembered enough to be useful.
It can be hacked. Identity verification systems are high-value targets because they gather exactly the information criminals want: names, dates of birth, document numbers, biometric signals, and account links. That is not ordinary web exhaust. It is the raw material of impersonation, fraud, account takeover, and identity theft.
It will be hacked. The only honest question is when, where, and how much damage follows. This is not theoretical. In October 2025, Discord disclosed that a third-party customer-service provider had been compromised and that government ID images connected to age-related appeals had been accessed by an unauthorized party. Discord later said approximately 70,000 users may have had government-ID photos exposed. The door had been convinced to speak: the ID photographs of tens of thousands of people may have passed into criminal hands.13
Why Verify the Prey?
Here is the question that the policy conversation has not yet answered satisfactorily: if the platforms are the maze and the business model is the Minotaur, why is the first serious answer so often to verify the prey rather than restrain the predator?
Meta, for its part, has publicly promoted app-store age verification in Canada, arguing that app stores should verify a user’s age before downloads and framing age assurance as an industry-wide youth-safety problem.14 A company that has spent years collecting behavioural data on users of all ages is now an enthusiastic advocate for age-assurance infrastructure, so long as the burden is shifted toward app stores and away from individual platforms like Facebook and Instagram. Meta seems to understand that the door is dangerous; it wants it built upstream, where someone else absorbs the risk when it fails. That is the tell: the risk has been moved, not solved.
There is another way to explore this story, one in which the maze is made answerable instead of every traveller at the door. It could mean banning targeted advertising to minors. Restricting compulsive design features. Requiring chronological feeds by default for young accounts. Limiting algorithmic amplification. Creating real liability for harmful design.
These measures cut into the maze itself rather than asking every traveller to empty their pockets at the door. They do not require the door to remember anything.
Manitoba has the right instinct: the Minotaur is real, and children have been sent into its house for too long. But the door being proposed is not only a shield. It is infrastructure that identifies, classifies, and retains. Infrastructure that can be breached, subpoenaed, repurposed, and linked. It will outlast the policy debate that created it and serve purposes no one in this conversation has yet imagined.
The Witness at the Threshold
Before Manitoba builds that door, there are questions worth asking loudly.
Who owns the memory the door creates? Who can compel it to speak: courts, agencies, governments, or the future owners of the vendor that processes the checks? When it is breached, what happens to the people whose information spills? What happens when the database built for a child’s safety becomes a liability they carry for the rest of their life?
A gate can keep danger out. It can also create a witness at the threshold: one that remembers every traveller who passed, what they proved, and where they were trying to go.
The child at the glowing door deserves better than a choice between the Minotaur and the witness. But if Manitoba builds the door it is proposing, the witness will remain long after the policy debate has moved on, keeping its quiet ledger of who came, what they proved, and which threshold they approached. That is not protection. That is a different kind of maze.
Geoffrey Dice is a writer based in Vancouver. This essay is Part I of The Age Gate and the Minotaur. Part II, Starve the Minotaur, is forthcoming.
Notes
- See Celia B. Fisher, Xiangyu Tao, and Madeline Ford, “Social media: A double-edged sword for LGBTQ+ youth,” Computers in Human Behavior 156 (2024); “Online Experiences and Mental Health of LGBTQ+ Young People.”↩
- See Global News, “Manitoba set to become 1st province to ban social media for children” and “Manitoba could fine social media companies billions if they don’t comply with youth ban.”↩
- See eSafety Commissioner, “Social media age restrictions.”↩
- See Global News, “Exclusive: Ontario considering social media ban” and Global News, “Manitoba set to become 1st province to ban social media for children.”↩
- See Reuters, “From Australia to Europe, countries move to curb children’s social media access.”↩
- See Reuters, “EU urges fast rollout of age-verification app to protect minors online,” European Commission, “Commission urges fast rollout of age verification app,” and Associated Press, “Roblox to require facial scans for under-16 users in Indonesia under new social media rules.”↩
- See U.S. Surgeon General, “Social Media and Youth Mental Health,” Common Sense Media, “AI Companions Decoded,” and Common Sense Media, “Talk, Trust and Trade-Offs: How and Why Teens Use AI Companions.”↩
- See Encyclopaedia Britannica, “Minotaur.”↩
- See Federal Trade Commission, “FTC Staff Report Finds Large Social Media and Video Streaming Companies Have Engaged in Vast Surveillance of Users with Lax Privacy Controls and Inadequate Safeguards for Kids and Teens.”↩
- See IAB, “Internet Advertising Revenue Report: Full Year 2025,” and Meta, “Meta Reports Fourth Quarter and Full Year 2025 Results.”↩
- See Office of the Privacy Commissioner of Canada, “Age Assurance -- Policy Note.”↩
- See “Habakkuk 2:11,” BibleGateway.↩
- See Discord, “Update on a Security Incident Involving Third-Party Customer Service,” and The Guardian, “Hack of age verification firm may have exposed 70,000 Discord users’ ID photos.”↩
- See Meta, “Canadian Parents Support App Store Age Verification for Teens When Downloading Apps,” and CityNews, “Meta pitching app-store age verification law to Liberal government.”↩