A little why can open a world.
A child notices the Moon through a car window. A curled leaf becomes a mystery. Three seeds become a first counting game. Curiosity gives ordinary moments somewhere to go.
AI creates new possibilities for exploring those questions. Yet families should not have to surrender their influence over a child’s learning environment, privacy, attention or exposure to unwanted influences to benefit from it.
Our mission: to open a world of knowledge and discovery to every child, while keeping parents at the center of their learning journey.
TrustinJr is being built as parent-guided AI discovery for early childhood, initially ages 4–5. Questions, prepared discoveries, illustrated stories and early math are different starting points. The child experience is audio-first on Apple Watch, with a parent experience on iPhone. The ambition is understanding that carries into life beyond the device.
Every child is the reach of our mission, not a claim about current access. The app is in development and testing, with device and onboarding limits. You can try the prepared concept now. For a practical introduction to evaluating tools like this, start with our guide for parents of young children.
Eight principles for the work.
Let curiosity lead.
A child’s question is an invitation to explore. Prepared discoveries can offer a spark when a child does not know what to ask. We want choices that feel understandable and worthwhile, with room for a child to change direction or stop.
Give a foothold. Leave room to think.
Scaffolding means offering enough support for the next step: a short explanation, a concrete example, a helpful clue or another attempt. The aim is to help a child do more of the thinking. More answers, longer sessions and correct taps do not, by themselves, demonstrate understanding.
Make the world part of the lesson.
A useful discovery can become something to notice, count, build, draw or discuss. Activities should have natural endings and invitations to explore with a grown-up. Physical play, books, rest, human teaching and caring relationships remain essential. A device should leave space for them.
Keep parents at the center.
Parents know a child’s readiness, history and family values in ways a tool cannot. The Parent Dock is intended to make experiences understandable and family preferences useful. Participation also means listening together, asking what a child thinks and helping check an uncertain answer. Visibility is support for that relationship.
Family preferences add to essential safeguards.
A family may prefer to explore a topic together. That choice is different from mandatory protections against exploitation, dangerous instructions or manipulation. Our design requires those protections to remain independent of optional preferences. A parent setting must not make harmful content permissible.
When safety is at issue, a child needs appropriate human help. No automated safeguard can promise to catch everything. The website’s sample topic switch demonstrates one scripted response; it does not verify the native app’s safety system.
Be clear about what AI is.
Trustin is a character in a computer tool. AI can give an answer that sounds confident and is wrong. Important, sensitive or confusing questions belong with trusted adults. The experience must never encourage secrets, emotional dependence or guilt about leaving. Parents, teachers and friends hold roles a learning tool cannot take.
Never sell a child’s attention.
No advertising to children, sponsored answers or purchase-pressure mechanics are operational commitments. So are finite activities, no endless content feed and no rewards for staying longer. TrustinJr should earn a family’s trust through its usefulness. Pricing has not been announced here.
Earn trust through verifiable choices.
Privacy-conscious design needs specific decisions about what is collected, why, who can access it and when it is deleted. We must explain those decisions plainly and verify them in the product. A promise, an implemented feature and an independently demonstrated benefit are different things. We will describe them that way.
One seed. A few thoughtful steps.
Imagine a child asking, “What is inside a seed?” A useful answer gives them an idea they can carry into their next observation.
| Moment | Illustrative response |
|---|---|
| Explain one idea | A seed holds a tiny young plant and food to help it get started. |
| Invite thinking | What do you think might start growing first? |
| Offer a helpful clue | A little root can start reaching out. |
| Connect to life | With a grown-up, look at a seed and a plant. What do you notice? |
| Leave space | Finish here. Let the next discovery happen away from the device. |
This illustrates our learning approach. It does not claim the current app assesses understanding or automatically delivers this whole exchange. Our scripted discovery and math samples let you inspect prepared content and a helpful retry. Personalized learning needs separate evaluation.
Research informs the work. Results must be earned.
The scaffolding tradition [1] and educational-app framework [3] give us useful design questions: Is the child thinking? Does the idea connect to something meaningful? Is there an opportunity to participate with another person? A framework helps us ask better questions; it does not establish that our answers work.
The guided-play review [2] examined children ages 1–8 and found benefits on selected outcomes. It excluded computer-guided studies. It therefore cannot establish the effectiveness of AI-guided learning. Research on responsive adult relationships [4] also cannot be used to portray a computer exchange as equivalent to human care.
UNICEF’s guidance [5] concerns children’s rights, including safety, privacy and accountability. AAP’s 2026 guidance [6] addresses the wider digital environment and includes early childhood. These inform our design responsibilities; neither organization endorses TrustinJr.
Common Sense Media’s 0–8 family research [7] helps us understand parents’ mixed hopes and concerns. It measures parent reports, not learning gains. Findings about teenagers are not evidence about four- or five-year-olds.
TrustinJr’s learning outcomes have not been independently validated. Enjoyment, parent feedback, completed activities and demonstrated learning are different evidence. We do not promise gains we have not measured.
What you can evaluate today.
TrustinJr is in development and testing. This website uses original artwork, prepared discoveries, a complete story, small math activities and recorded narration. Its browser interface is adapted for demonstration. The illustrative Parent Dock changes this demo only; no real child profile, live AI conversation or microphone recording is involved.
The current beta requires an Apple Watch normally paired with an iPhone. Family Setup onboarding remains unresolved for that beta. Public App Store availability, release timing, pricing and broader age progression are not confirmed here. Check current availability before buying a device.
Privacy is a separate product responsibility.
The website privacy notice describes this site. Native-app data handling requires its own verified disclosures, retention rules, access controls and appropriate consent process before public launch. No-advertising design alone does not establish privacy compliance. We do not claim zero data processing, certification, perfect filtering or guaranteed accuracy.
What should be evaluated next?
- Whether children and parents understand the experience and its limits.
- Answer accuracy, age suitability and the usefulness of each learning step.
- Control reliability when audio, connectivity or family preferences change.
- Mandatory safeguards, privacy practices and family settings as separate responsibilities.
- Whether a child can apply an idea in a new task or real-world observation.
Read the parent starting guide or explore both sides of the prepared preview.
References and their limits.
- Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Foundational work on human tutoring and scaffolding; not an evaluation of AI or this product.
- Skene, K., et al. (2022). Guided play: systematic review and meta-analysis. Review of 39 studies; 17 in the meta-analysis. Children ages 1–8; adult guidance only. Outcomes and study quality varied.
- Hirsh-Pasek, K., et al. (2015). Putting Education in “Educational” Apps. Research-informed design framework for children’s apps; not a trial of TrustinJr or a certification standard.
- Harvard Center on the Developing Child. Serve and Return. Explanation of responsive relationships between young children and caring adults; not evidence that AI can supply an equivalent relationship.
- UNICEF (December 2025). Guidance on AI and Children, version 3.0. Child-rights policy and system-design guidance across childhood; not preschool learning-outcome research.
- American Academy of Pediatrics (January 2026). New digital media recommendations. Current guidance replacing prior AAP media statements; distinguishes early childhood 0–5, school age and adolescence.
- Common Sense Census: Media Use by Kids Zero to Eight (2025). Probability survey of 1,578 U.S. parents, August 2024. Age bands include 2–4 and 5–8; no distinct 4–5 outcome study.
Sources reviewed October 13, 2026. References explain what informs the approach and do not imply affiliation, endorsement or certification.
1.2 · October 13, 2026 — Mission-led revision, three brand promises, parent participation, clearer evidence populations and linked parent guide.
1.1 · October 12, 2026 — Added a parent-friendly summary.
1.0 · October 12, 2026 — Initial philosophy and product-stage disclosure.
Questions or corrections: [email protected].