Computerphile
June 16, 2026
Technical Challenges of Banning Social Media for Under-16s
TL;DR
A technical analysis of how governments could implement age-based social media bans, examining blocking methods at routers, ISPs, devices, and platform level, and exploring the feasibility and tradeoffs of each approach.
Key Takeaways
- 1Blocking at home router or ISP level is impractical because children use multiple networks and devices, requiring universal age-verification protocols that don't yet exist
- 2Device-level controls (like Apple Screen Time or Google Family Link) require accurate age data entry, which is easily circumvented by users entering false birthdates
- 3Platform-level enforcement places responsibility on social media companies to verify age, using methods like government ID APIs, facial recognition, credit cards, or third-party verifiers like Yoti
- 4No age-verification system is perfect—determined users can bypass checks using VPNs, fake documents, or borrowed accounts—but modest barriers may shift societal norms over time
- 5Facial recognition for age verification is unreliable due to variable child development rates and AI model limitations, as demonstrated by Australia's experience with fake mustaches defeating checks
- 6Alternative approaches include credit card verification (which teenagers can access) and phone OS providers performing age checks, but all carry privacy and circumvention risks
- 7The most viable model mirrors drink-driving legislation: imperfect enforcement combined with cultural shift, aiming to normalize avoiding platforms until age 16 rather than achieve perfect compliance
Notable Quotes
“If you're determined to lie about your age, to steal someone's credit card, right, to put a fake mustache on, yes, you will get around these things.”
— Host
“The only perfect system would be everyone has some kind of barcode tattooed on their hand and you know there's someone sitting over your shoulder all the time.”
— Host
“Maybe if you make YouTube Kids as good as YouTube for kids, then they don't even care about going on YouTube, right? That would be the biggest win for me, I think.”
— Host
“I think as a society we don't think children are well served by social media right now and I don't think the social media companies have been seen to be particularly proactive in doing anything about this now they have to.”
— Host
Chapters
1. The Problem and Government Proposal
Governments propose banning under-16s from major social media platforms like TikTok due to addictive design and problematic content. The technical challenge is implementing such a ban across diverse internet access methods.
2. Network Architecture and Blocking Points
Home networks, ISPs, and mobile providers form a chain from device to server. Blocks could theoretically occur at the router, ISP, or phone company level, but each has significant practical limitations.
3. Home Router and ISP-Level Blocking
Technologies like Pi Hole can block DNS requests, but implementation requires technical expertise beyond average users. ISP-level blocks fail because children access the internet from multiple networks (friends' homes, cafés) where others shouldn't be restricted.
4. Device-Level Controls
Apple Screen Time and Google Family Link allow parent control, but rely on accurate age data during phone setup. Users easily circumvent by entering false birthdates, and concentrating gatekeeping power in two companies raises governance concerns.
5. Platform-Level Age Verification Methods
Government regulation targets social media companies directly. Proposed verification methods include government ID APIs, facial recognition AI, credit card checks, and third-party verifiers like Yoti, each with distinct tradeoffs.
6. Government ID and Privacy Concerns
Digital ID verification is technically feasible and most secure, but public and policy resistance cites privacy concerns. Users could impersonate others by entering false credentials despite API checks.
7. Facial Recognition Limitations
AI age-detection models are not 100% accurate and struggle with rapid child development variation. Australia's experience showed fake mustaches defeating facial recognition systems, demonstrating insufficient rigor.
8. Alternative Verification: Credit Cards and Phone Providers
Credit card requirements assume age thresholds, but teenagers can access cards via parents or fintech apps. Phone OS providers (Apple/Google) already know device age and could verify without extra steps, but inherit the same falsification risks.
9. Accepting Imperfection and Cultural Shift
No system is perfect; determined users will circumvent via VPNs, fake documents, or borrowed accounts. The model mirrors drink-driving policy: gradual cultural normalization that discourages most people without catching everyone.
10. Opportunity for Safe Alternatives
Rather than only restricting platforms, companies should develop age-appropriate versions (e.g., YouTube Kids improvements). Strong parent support (~90%) and international adoption suggest this approach will persist despite technical imperfection.
Key People & Entities
- Yoti
- Third-party age verification company used in Australia and proposed for UK regulation; stores identification data and issues challenge tokens to confirm user age
- Apple (Screen Time) and Google (Family Link)
- Operating system providers offering parental control features that can restrict app access based on device-level age data
- TikTok, Meta, Snapchat
- Major social media platforms targeted by government age-restriction legislation
- YouTube Kids
- Age-appropriate alternative to YouTube, exempt from proposed restrictions; cited as a model for safe platform alternatives
Glossary
- DNS (Domain Name Server)
- A system that translates domain names (e.g., tiktok.com) into IP addresses; can be configured to block access to specific sites by returning 'no such address' responses
- Pi Hole
- A DNS filtering tool that blocks advertisements and specified websites at the router level; commonly used for content filtering but requires technical setup
- VPN (Virtual Private Network)
- A technology that masks a user's IP address and location, allowing circumvention of geographic or network-level blocks
- Convolutional Neural Networks (CNNs)
- AI models used for image recognition tasks, including facial recognition and age estimation from photos
- Facial Recognition for Age Detection
- AI models trained to estimate a person's age from a photograph; limited in accuracy due to variable human development rates
- Third-Party Age Verifier
- A company (e.g., Yoti) that stores identification data and provides challenge tokens to platforms confirming a user's age without platform companies holding the data directly
- Challenge Token
- A digital credential issued by a third-party verifier confirming that a user meets age requirements, without revealing exact age or identity details to the platform
- ISP (Internet Service Provider)
- The company that provides internet access to homes and devices; potential point for network-level content blocking