A Regulator Announces a Plan, Then Immediately Hedges It
On July 16, 2026, Kim Jong-cheol, chairman of South Korea's Korea Media and Communications Commission (KMCC — renamed from the Korea Communications Commission in October 2025), told a presidential policy briefing that his agency is "reviewing, in phases, a plan to restrict children under 14 years of age from creating accounts on social media platforms," and separately looking at limits on "designs and algorithms that could encourage excessive use" for users aged 14 to 19 (Korea Times). The vehicle would be a revision to the Act on Promotion of Information and Communications Network Utilisation and Information Protection — commonly called the Information and Communications Network Act — a bill amending that statute was already introduced in the National Assembly on June 22, 2026, citing a rise in cases of harm to children and adolescents online (Digital Policy Alert bill tracker).
What's notable is what Kim said next: that the government must move carefully, pointing explicitly to the 2011 "Shutdown Law," which barred under-16s from online games between midnight and 6 a.m. and was repealed in January 2022 after a decade of evidence that it barely worked and mainly generated resentment. Seven separate youth-SNS bills are already pending in the National Assembly, several of them stalled for months, and Kim has said repeatedly that legislating this is properly the Assembly's job, not his agency's — the KMCC's role is to "prepare policy alternatives and support public discourse" (Money Today; Korea JoongAng Daily). That is an unusually candid admission from a regulator floating a headline-grabbing restriction: the last time Korea tried a blunt-instrument age rule for online life, it didn't survive contact with reality.
The Steelman: The Numbers Are Genuinely Bad
The case for doing something is not manufactured. South Korea's Ministry of Gender Equality and Family's 2025 Youth Statistics found 42.6% of teenagers (ages 10–19) fall into the smartphone over-dependence risk group, up 2.5 percentage points from 2023, with middle schoolers the most exposed at 41.7% (Kukmin Ilbo, citing the ministry's release). That is nearly double the 22.7% risk-group rate across the general population reported in the government's own composite indicator dashboard (index.go.kr, Ministry of Science and ICT data). Korean researchers and regulators point to a specific mechanism — infinite scroll and engagement-optimized recommendation feeds — that Kim's proposal explicitly names as the target for 14–19 year olds, not blanket use restrictions. A regulator narrowing its focus to addictive design, rather than banning a platform category outright, deserves credit for at least attempting proportionality rather than a blunt ban.
Why the Under-16 Precedent Should Make Seoul More Cautious, Not Less
The obvious model is Australia's under-16 social media account ban, in force since December 10, 2025. Its early results are a warning, not a template. Australia's eSafety Commissioner's first compliance report found that more than 80% of Australians under 16 are still using social media, that 81% of 10–15 year-olds kept or created accounts within three months of the ban taking effect, and that daily use barely moved — 60% used social media daily before the ban, 58% after (Al Jazeera, Aug. 3, 2026). About half the kids who kept accounts said no one ever checked their age; others simply lied about being 16, or beat face-scan verification. Meta reports removing over 750,000 under-16 accounts in Australia — real numbers, but dwarfed by the share of minors who stayed on regardless. New Zealand introduced its own under-16 account bill on August 24, 2026, following the same design logic — personalized recommendations plus continuously updated feeds as the triggering features (MediaNama) — meaning three democracies are now running variations of the same experiment roughly in parallel, with the earliest data point already showing the enforcement gap.
Age-verification technology itself doesn't close that gap. EFF's review of newer approaches, including zero-knowledge-proof age credentials, concluded in August 2026 that ZKPs "aren't a silver bullet" — they can reduce some privacy exposure relative to ID uploads or face scans, but they still require a trusted issuer, still can be spoofed or shared, and still don't solve the underlying problem that a 13-year-old with a parent's ID or an older sibling's help defeats almost any check (EFF). Korea would be adopting a mechanism whose flagship implementation is, eight months in, catching a minority of the population it targets.
The Free-Expression Backdrop Matters Here Too
This proposal doesn't arrive in a vacuum. The same Information and Communications Network Act was amended in December 2025 to target disinformation, passing the Assembly 170–3 with new liability for creators who spread "illegal, false or manipulated information." A month later, US Under Secretary of State Sarah Rogers publicly warned that the revision handed regulators an "invasive license for viewpoint-based censorship" and could "endanger" US-Korea tech cooperation, urging civil remedies over regulatory discretion instead (Korea Times). Layering a second major youth-safety mandate onto the same statute, administered by the same agency, invites the same scrutiny — especially since algorithm restrictions necessarily require regulators to define what counts as an impermissibly "addictive" design choice, a judgment call adjacent to content policy.
The Better Path Is the One Kim Is Already Describing
Credit where due: Kim's own framing — phased, evidence-based, routed through the Assembly rather than administrative fiat, explicitly wary of repeating the Shutdown Law's failure — is closer to sound policy than most of the peer proposals it's being compared to. The mistake would be treating Australia's account ban as proof of concept rather than as the live, imperfect experiment it currently is. A narrower rule focused on default settings — disabling autoplay and infinite scroll for minors' accounts unless a parent opts in, rather than gating account creation on unreliable age checks — would target the actual harm (engagement-optimized design) without reproducing a verification regime that evidence to date shows is easy to route around and hard to enforce without intrusive data collection. Korea has the rare advantage of watching Australia's ban fail in close to real time before writing its own law. It should use it.