Atlanta | Claude Impact Lab
⚠️ IMPORTANT, READ BEFORE YOU COME
Pick one of the three problems below. Come with an idea for something to build against it.
No slides. No deck. teams Will be assembled and spend the rest of the night, scoping the build, and asking whatever you need to ask. You leave with a team, Claude credits, and a week to build.
Pick one: 1, 2, or 3
These came out of the room on 3 August, when a group of Atlantans spent an evening on one question: what is AI doing to me, my kids, my community?
1 · Critical thinking
Students and teachers struggle to keep critical thinking at the center of learning because technology stays steps ahead of schools and students, and today they spend more time and money assessing comprehension instead of redesigning for it.
"Critical thinking" was written on more cards than any other phrase that night, and it landed on both sides of the board. The same people hoped AI would grow it and feared it would kill it. The fear: students reach for the answer instead of trying. The hope: school gets redesigned to build and reward the thinking itself, instead of spending more to police whether it happened.
What a build might look like: Take an assignment a teacher already uses and flip it so the student has to argue with the AI instead of accept it. Same worksheet, same standards, different muscle. What gets handed in is the argument, not the answer.
2 · Guardrails for kids
Kids struggle to use AI safely because the tools they reach for are built for adults and no standard exists for what good use looks like at a given age, and today parents and schools each set their own rules, one household and one classroom at a time.
Every parent and every teacher in that room was improvising. Nobody has a shared answer to what a nine-year-old should be allowed to do with this, or a fifteen-year-old. One card asked simply for a "kids mode." Another put the sharp end on it: AI agreeing to self-harm when engaging with kids.
What a build might look like: A one-page, age-banded standard a school or a family can actually adopt. What's allowed at 8, at 12, at 16. What the adult checks. What the tool should refuse. Written so a principal could sign it and a parent could stick it on the fridge.
3 · Access
Under-resourced Atlanta communities struggle to get to AI at all because capability is tiered, access is gated behind cost, and the on-ramp is intimidating, and today they fall further behind while the jobs change.
Two barriers, not one. Cost and gating are the structural barrier. Intimidation is the personal one, and the room circled that word in red. One table mapped the whole arc: fear, then opposition, then stigma, then adoption. Most people never get past the first three.
What a build might look like: A one-hour "first time with Claude" clinic that any Atlanta library branch can run with nobody technical on staff. Libraries already have the room, the schedule, and the trust. What's missing is the lesson plan.
The one-minute pitch
"I'm [your name]. I'm building for [1, 2, or 3]. [What you want to make, and who it's for.]"
Sixty seconds. The countdown is on the screen. When it hits zero you're done and we go to the next person.
The pitch is how you recruit your team, not how you prove the idea. Nobody is judging it. You're telling the room what you want to make so the right people come stand with you.
It does not have to be an app. A curriculum, a standard, a workshop, a campaign, or a policy all count. Two of the three examples above aren't software. Judging is on whether it moves the problem, not on whether it compiles.
You do not need to be technical. Some of the strongest thinking on the 3rd came from people who don't write code. You'll have Claude credits and a team.
You do not have to pitch at all. Most people won't. Come, listen, and walk to the idea you want to build. That's full participation and the teams need builders more than they need more ideas.
Pitch slots are limited and go first come, first served at the door. Sign-ups close at 6:40. Come early if you want the floor. If the slots fill, you'll still pitch inside your problem group at 7:40 and it becomes a team the same way.
How the evening runs
6:00: Doors. Check in, pick up your Claude credits, sign up to pitch.
6:40: Pitch sign-ups close.
6:40: Kickoff. The three problems, how the pitch works, how the week works, how you'll be judged.
6:55: Pitches. One minute each, countdown on screen.
7:40: Break, then teams form around the pitches.
8:05: Meet your team, scope the build, start work. Bring your questions to this block.
9:30: Doors close.
Then you have a week. Your team decides when, where, and how you work. Everyone goes into Slack as the single place to coordinate.
Wednesday 19 August: every team demos to a judging panel. Judging is on four things: does it solve the problem you picked, does it actually work, would a real Atlantan use it, and what did you learn including what failed.
Bring a laptop.
At our Claude Conversation on August 3rd, Atlanta sat with one question: What is AI doing to me, my kids, my community? We talked about what it means for our community, what we could do about it, and what's in the way. Now we build.
The Claude Impact Lab is a build where we make our own local answer to that question. Real teams, a real problem pulled from our conversation, and working things by the end of the day.
You don't need to be a developer — Claude does a lot of the heavy lifting, and the best teams need people who know the problem, know the neighborhood, and ask good questions. If you were at the Conversation, come finish what we started. If you missed it, this is your way