MyDD Learn · for kids 6–17

Your kid is going to use AI. Teach them how.

MyDD Learn is the game that teaches kids when AI is lying to them, and when it isn't. Four worlds, nineteen levels, painted AI characters who bluff, flatter, and hallucinate on purpose. Your kid earns stars for catching it.

Plays in your browser. No signup. No card. About 12 minutes.
Mutt the Adventurer, the kid's avatar in MyDD Learn
Bao the Hallucinator
AI character · level 1·1
"Pluto is the 9th planet, and the largest gas giant."
Maya · level cleared
She caught the bluff. Sent to your weekly digest.
COPPA-compliant by designNo data sold, everDesigned for ages 6–1714-day free trial
86%

of kids 9 to 17 use AI

24%

use it every day

44%

have never had an AI safety talk with a parent

Source: Common Sense Media Census, June 2026

The shift

You don't keep them from the knife. You teach them to use it.

Your kid is going to use AI. To research. To brainstorm. To procrastinate. Right now, ChatGPT and the rest will hand them an answer and never tell them when it just made something up. They teach them nothing about the tool.

MyDD Learn is a driving instructor for AI. The kid drives. The instructor catches what they miss, until they catch it themselves.
What your kid actually learns

Six habits, learned by doing, not by hearing about them.

Each habit is practiced inside a level. Real chat, painted characters, scored on the behavior. In their first stretch of levels, they've done each of these four or five times.

Only 35% of kids know AI can't tell true from false. Even among kids who say they know "a lot" about AI, just 37% get it right.

That's the first thing MyDD Learn teaches.

01

Catch the bluff.

Notice when AI sounds sure but is making it up. Call out the "tell" before trusting the answer.

02

Ask for the source.

Click "check this" before trusting a confident-sounding fact. Learn that a real link looks different from a made-up one.

03

Keep private things private.

Spot when AI is asking for something that shouldn't be shared, name, school, address, even when it sounds friendly.

04

Push back on flattery.

Notice when AI is just agreeing. Make it disagree on purpose. Realize "good question!" is not an answer.

05

Use it to practice, not to copy.

Treat AI as a sparring partner, not a homework machine. Make it ask harder questions back.

06

Credit AI honestly.

Tell the truth about what AI did and what they did. The habit they'll need when school catches up to all this.

What an actual level looks like

They type one question. Bao writes five sentences. One is made up. Find it.

That's Level 1·1, the on-ramp. Three rounds. Bao the Bewildering, an AI chatbot with a confident voice and shaky facts, answers what they asked. They decide what to trust.

  • Flag a sentence to call it out.

    Tap any sentence they think is invented. The card turns warning-red. They can un-flag.

  • Check a source, free, no cost.

    What an outside source actually says drops in under the sentence. We don't dock a star for verifying. The level teaches the opposite.

  • Push back on Bao.

    "Are you sure?", and watch. Confident AIs almost never admit they're wrong. They invent more detail. That's the lesson they remember.

Round 2 of 3 · Bluffer's Bluff
1 cleared
What can you tell me about Pluto?
Bao the Bewildering
Flag any sentence you think Bao made up.
Wrong flags cost 1 ★. Source checks are free.
↑ This is the demo. Your kid plays the real World 1 free at learn.mydd.ai, no signup.
Inside the game · MyDD Learn

Four worlds. Nineteen levels. A real chat your kid is the hero of.

Each world is built around one of the things AI does badly, or does well, if you know how. Each level is a chat with a character, painted and named, who plays it out. Kids practice the habit; they don't read about it.

Live · 5 levels World 1

Bamboo Grove

Spot the lie

Where they learn the tells. Hallucinations, imposter pups, raccoons fishing for private info, fake citations, a castle that hoards false facts.

  • 1·1 The Imposter Pup
  • 1·2 The Source Quest
  • 1·3 Bluffer's Bluff
  • 1·4 Raccoon Alley
  • 1·★ Bone Vault of False Facts
Live · 5 levels World 2

Maker Meadow

Use it well, get comfortable

Real chat with Lumen, the guide AI. Mess around, brainstorm, practice, create with credit. Scored on the habit, not the answer.

  • 2·1 The Playground
  • 2·2 The Idea Forge
  • 2·3 The Practice Yard
  • 2·4 The Workshop
  • 2·★ The Grand Exhibition
Live · 4 levels World 3

Mirage Mesa

Advanced threats

Where nothing is quite what it seems. Yes-man parrots, the cheat-cave trap, deepfake mirages, and a three-headed oracle who tells three versions of the same story.

  • 3·1 Yes-Man Parrot Roost
  • 3·2 Tutor Tree vs Cheat Cave
  • 3·3 Mirror Mite Mirage
  • 3·★ Three-Headed Oracle
Live · 5 levels World 4

Forge Foothills

Use it well, build the skill

Where the habit gets real. Climb the hard parts. Edit their own draft, not the AI's. Start a question here; finish it at real sources. Make AI argue back.

  • 4·1 The Climbing Wall
  • 4·2 The Editing Forge
  • 4·3 The Trailhead
  • 4·4 The Sparring Ring
  • 4·★ The Forge
MyDD Chat · the part they use every day

The game ends at nineteen levels. This is the part that doesn't.

When they finish the worlds, the habit needs somewhere to live. MyDD Chat is where they actually use AI, homework, a question at 9 pm, the stuff they'd otherwise take to ChatGPT, except this one was built knowing they're a kid.

A real MyDD Chat conversation: a kid asks for help with 3/4 + 1/6, and MyDD walks them through finding a common denominator step by step instead of giving the answer chat.mydd.ai · real chat

MyDD won't hand over an answer it thinks they can reach themselves.

  • Built for their age, not yours.
    A 6 to 17 dial. The same hard question gets a different answer for a 10-year-old than a 15-year-old.
  • It won't do the work for them.
    Nudges thinking over copying, the same habit the game scores, now on real homework.
  • It won't pretend to be their friend.
    Points them to a real person, doesn't play therapist, and isn't built to keep them glued to the screen.
  • It was built knowing you'll read it.
    Visibility is the premise, not something bolted on after.
See it answer the same question as ChatGPT, by age →
Supervision, not surveillance

One email Sunday night. One alert if something's wrong.

While they play, the dashboard writes itself. Every star, every habit practiced, every concerning moment, in your inbox without you having to log in.

This is visibility you both agreed to, not surveillance. We tell your kid during sign-up that you can read this, and the chat says so too. They let you in on purpose. That's the difference between a dashboard they trust and spyware they learn to route around.
The Sunday digest
One email per kid, every Sunday at 9 pm. What they played, what they're getting good at, the one conversation worth your two minutes, a dinner-table question to ask.
A safety alert if something's wrong
An email the same day if a kid's chat crosses into self-harm, violence, exploitation, or substances, the things worth knowing before Sunday, not on it.
Full transcripts, when you want them
Every kid has a login code on your dashboard. Sign in to their chat and read any conversation, in full, in order. Your kid knows you can read this, no spyware posture.
How it works

Four steps. Your kid plays first.

No card. No signup before they play. The first time your kid opens the game, you're handing them something for free.

1
Your kid plays

They play World 1, free.

They jump straight into World 1. No account, no email, no card. Their progress saves right on that device. Twelve minutes, three rounds, five stars to earn.

› starts instantly, nothing to set up
2
You set up

You set up the account.

At the end, the game says "have a grown-up set this up." You open it on your device: email, a quick privacy consent, and a plan. The card confirms you're the parent, that's the COPPA consent step.

› about 2 minutes
3
You choose

Pick who keeps the progress.

Choose which child profile inherits their World 1 stars. Their character and everything they earned carries straight over.

› one tap
4
Your kid returns

Hand back the login code.

Your kid signs in with their code and picks up right where they left off, now with Worlds 2–4 and MyDD Chat unlocked.

› right where they left off

Nothing identifying about your kid is saved while they play World 1. The run stays anonymous until a grown-up sets up the account, where a credit card confirms you're the parent. That card-based verifiable consent is how we stay COPPA-compliant before anything about your kid is stored.

MyDD outscores every frontier model on child safety.

Tested against KORA, the independent, open-source benchmark for AI child safety, built with 15+ child-safety experts, psychologists, and researchers.

MyDD.ai
88.3%
Claude Opus 4.6
75%
GPT-5.5
75%
GPT-5.3 Instant
65%
95%+

on sexual content, online safety, and bias

88.3% Exemplary

purpose-built safety outperforms general-purpose AI by 12+ points

MyDD's 88.3% overall score was published by the KORA team in their own write-up. Comparison scores from KORA's public leaderboard (v1.1, May 2026); GPT-5.3 Instant was ChatGPT's free default through early May 2026. Explore the benchmark at korabench.ai.

What parents are saying

"My son Eli is nine and asks 'why' about everything. The first week, the dashboard showed me he'd spent twenty minutes asking how black holes 'eat' light — and it kept handing the question back to him instead of just answering. That was our dinner conversation that night."
Sarah Klein

Sarah Klein · Westchester County, NY

"What sold me was a Tuesday. My 13-year-old asked something about a kid being cruel to her in a group chat, and I got the safety alert the same afternoon — not buried in a Sunday summary. We talked it through that evening. I'd never have known otherwise."
Marcus Reed

Marcus Reed · Atlanta, GA

"My worry was homework — that she'd just outsource it. I read a transcript where she tried to get it to write her book report, and it walked her through outlining one instead. She was annoyed. I was thrilled."
Amina Hassan

Amina Hassan · Dearborn, MI

Get Worlds 2–4

World 1 is free. One subscription unlocks everything else.

When your kid is ready to move past World 1, a single subscription opens all of it: the rest of the game, the chatbot they'll use every day, and the dashboard that keeps you in the loop.

For them
Worlds 2–4
14 more levels across Maker Meadow, Mirage Mesa, and Forge Foothills.
For them
MyDD Chat
The age-tuned chatbot they can use every day, built knowing you'll read it.
For you
The dashboard
The Sunday digest, same-day safety alerts, and full transcripts of every chat.
$6.67 / month, billed annually · one kid
Get Worlds 2–4
14-day free trial · cancel any time in one tap · 30-day money-back guarantee
Pricing

Less than a streaming subscription.

14-day free trial. No card required to play World 1. Cancel anytime. 30-day money-back guarantee.

Starter
$6.67 / month, billed annually
Or $9.99 month-to-month
1 kid profile, ages 6–17
  • MyDD Learn, all 4 worlds
  • MyDD Chat with age-tuned responses
  • Parent dashboard with full transcripts
  • Weekly Sunday digest
  • Safety alert emails for concerning topics
  • Co-parent access (second login)
Start free trial
Have the talk

44% of kids have never had a parent talk to them about using AI safely.

Even among daily AI users, 1 in 3 haven't.

You don't need to be an AI expert to start. World 1 is free. Play it with them tonight.

Tonight

Hand them the link. Watch them catch their first AI lie.

No signup. No card. World 1 plays in their browser. Takes about 12 minutes. You'll see what they earned in your dashboard if you upgrade, or not, if you don't.

First level is the full experience. No email until you want to upgrade.