Why this exists
Children under five are spending their most formative years inside media their parents can no longer vet by hand. In 2025–26 the ground shifted:
- After watching a popular preschool show on YouTube, 40% of the next recommended videos were AI-generated (New York Times investigation, cited in the April 2026 Fairplay coalition letter to YouTube).
- The top AI-slop channels aimed at children earn over $4.25M/year — a permanent economic incentive to make more.
- The FTC fined the largest children's media company in the world $10M for mislabeling kids' videos and collecting children's data without consent (September 2025).
- 100% of subscription kids apps audited rate medium-to-high on privacy risk.
Parents are not asking for more content. They are asking a question no one answers honestly: is this one safe, calm, and made by a human who cares?
LUMI Certified Calm is our answer, written down, so it can be checked. We hold everything we make to it. We publish it openly so any maker can adopt it and any parent can hold us to it. A standard you can't verify is marketing; this one is built to be audited.
The five commitments
A piece of media — a song, a story, an activity, an app screen, a video — is LUMI Certified Calm only if it meets all five. Not four. All five.
1Calm by design
It does not hijack a developing nervous system to hold attention.
- Pace: no cut/scene change faster than the child can follow. For audio, tempo stays in the engaged-listening zone (78–115 BPM; never the 115+ agitation zone).
- Sound: no sudden loudness spikes, no alarming SFX, no "attention-grab" audio design. Volume is even; silence is allowed.
- Visuals: no strobe, no hyper-saturation, no relentless motion, no more than a calm number of things happening at once.
- The test: a two-year-old can look away without distress, and come back without having missed a cliff-hanger engineered to trap them. Calm content survives being paused.
Reference bar: passes the strict low-stim test that even Bluey narrowly fails (louder moments, quicker cuts). Our canon: the register of Puffin Rock, Sarah & Duck, and a parent's own voice.
2Made by a human who is accountable
A named person is responsible for every published piece, and stands behind it.
- Every asset has a human author or curator of record. Not a channel. A person.
- AI may accelerate production (drafting, voice synthesis with disclosure, translation-assist). AI may never be the author of record and is never presented to families as the maker.
- No templated, plotless, infinitely-generated content. If it could be produced a thousand times a day with no one reading the output, it fails.
- Disclosure, always: where synthetic voice or AI-assisted production is used, it is declared in metadata (DDEX synthetic-content flags for audio; a plain-language note otherwise). We never say "AI-generated" in family-facing copy — because it isn't; a human made the decisions. But we never hide the tools either.
This is the line between LUMI and the AI studios named in the Fairplay letter. Same tools; opposite accountability.
3Zero data on children
The child is never the product, the account, or the datapoint.
- No child accounts. No child profiles. No child data — ever. The parent is the only account holder.
- We measure content, never children: which songs get replayed, which activities get saved — aggregate, never individualised (the same disclosed posture as the screen-free audio leaders). We could not build a profile of your child if we wanted to, because we never collect the inputs.
- No behavioural advertising to children. No selling, sharing, or brokering of any family data. No third-party trackers on child-facing surfaces.
- Personalization (a child's name in a song) is rendered and not stored — the name exists for the moment of creation and then is gone.
Every year, regulation (COPPA amendments, UK OSA, EU DSA minor guidelines) raises the cost of collecting kids' data. We designed to zero, so the rising tide only lifts us.
4Pedagogically honest
It respects how a young child actually develops — and never fakes it.
- Age claims are truthful and specific (a 0–12m activity is genuinely for 0–12m, reviewed against developmental norms).
- Educational claims are grounded, not marketing. We do not promise a screen will make a baby smarter.
- Co-play over solo-consumption: our activities are built for a parent and child together, not to park a child alone.
- Cultural and linguistic content (including our Arabic line) is reviewed by a native-fluent, culturally-fluent human — never machine-translated and shipped.
5Parent-respecting, guilt-free
It treats the parent as an intelligent adult, not a target.
- No dark patterns: no fake urgency, no manipulative unsubscribes, no "your child will fall behind" fear-selling, no engagement-maximising notifications.
- The free thing is genuinely free and genuinely useful. The paid thing is honestly priced and easy to leave.
- We market against screen guilt, never with it. The goal is a calmer household, sometimes including less of us.
How we prove it (the receipts)
A standard no one can check is a slogan. Ours is auditable:
- 1The public ledger. A "How We Make Things" page describes our actual pipeline — human authors, review gates, disclosure practice, the zero-data architecture — in plain language, kept current.
- 2Per-asset provenance. Audio ships with DDEX flags; every published asset carries its human author of record and review date.
- 3The privacy architecture is inspectable. No child-data collection is not a promise we ask you to trust — it's visible in what the products don't ask for (no child name field that persists, no child login, no trackers).
- 4Independent alignment. We map this standard to the external frameworks that matter — AAP screen-time guidance, COPPA, UK OSA children's codes, EU DSA — and invite pediatric, ECE, and child-safety reviewers to hold us to it. (Standing goal: be cited by the advocacy organizations doing the work, not defensive against them.)
For other makers
This standard is open. Any creator of media for young children may adopt it, hold themselves to it publicly, and say so. We would rather live in a world of many calm, honest, privacy-safe makers than be the only one. A rising standard is a rising floor — and the floor is what protects the children.
If you make things for the first five years and you meet all five commitments, you are welcome to stand on this. Over time, LUMI intends to make "Certified Calm" a mark others can earn — a hub for the calm-media movement, not a fence around one company.
What this is not
- It is not a claim that screens are good for babies. The AAP is right: less is better, none is best under 18 months. When we make the least-bad option, or the thing that replaces a screen entirely (a song, an activity, an audio story), we are doing our job.
- It is not a legal certification or a regulator's seal. It is a public promise with public receipts, which is more than most of the industry offers and exactly what we can be held to.
- It is not final. Version 1.0. The bar only goes up.
LUMI is a calm, kind children's media house for the first five years — parent-made, produced with disclosed and invisible AI assistance, owned end-to-end, and native to both English and Arabic. Calm. Kind. Curious.