AI access & inequality

The most common worry people bring to this camp isn't whether AI works — it's who gets to use it. Frontier tools sit behind subscriptions; running models locally needs hardware; the people most lectured about "keeping up" are often the ones priced out. This page is where the camp says plainly what it does about that, and points to the parts of the site that hold the conversation up.

§ A · The gap, named

cost, hardware, and the helplessness around both

"Just use AI" assumes a credit card, a recent laptop, reliable internet, and the time to learn — four things that are unevenly distributed by design, not by accident. A teacher in an underfunded school, a student between jobs, an artist in a country where a single month of a frontier subscription is a meaningful sum: each is told the tools are essential and that access is their own problem to solve.

The camp treats this as a first-order ethics question, not a footnote. If a workshop about AI can only be done by people who already pay for AI, it has quietly chosen its audience. So the camp is built to be done with nothing but a browser — and the access conversation itself becomes one of the things we investigate.

Try it as an investigation: the Access Tiers tool runs one classroom task three ways — paid frontier tool, free limited tool, and unplugged or local alternative — so the cost of access becomes something you can see in the output, not just argue about.

§ B · What the camp does by design

access is built in, not bolted on

The camp runs on a browser and nothing else

Every core teaching moment uses simulations and frozen examples, never a live account — so no participant ever has to pay, log in, or own hardware to do the whole camp. That is a deliberate equity decision before it is a reliability one.

No-AI pathwayA full route through every session without direct generation — first-class, not a fallback Recordings & async routeThe whole camp at your own pace — no live attendance, no premium tools required Access Tiers toolCompare the same task across paid, free, and unplugged tiers Younger-learners adaptationsAccount-free by construction — children never need live AI

§ C · Free & low-cost ways in

when you do want to go further

The belief that real AI work requires an expensive subscription is widespread and increasingly out of date. You can train a model in the browser for free, run capable open-weight models on a normal laptop, and call frontier models through free API tiers — none of it requiring a paid plan. The Hands-On page lays out the specific tools and exactly what each one needs.

Train in the browser, freeTeachable Machine, ML for Kids, AI for Oceans — no account, no install Run models on your own machineOllama & LM Studio — open-weight models, no subscription, nothing sent anywhere APIs on a free tierFree quotas and pay-per-token pricing — pennies, not a monthly fee Library, not bookstoreThe reading list is built around free essays and books in most library systems

For the classroom specifically: "free" and "safe for students" are not the same test. Everything in this camp runs in the browser with no logins, so nothing about a student is collected — that's the safest default. When you do reach past it, prefer tools that are account-free, age-appropriate, and don't retain student input: Teachable Machine, Machine Learning for Kids, and AI for Oceans all run without student accounts, and the younger-learners adaptations are account-free by construction. Always check a tool's data-retention and minimum-age terms before a whole class uses it.

§ D · Read the politics of access

free/local is a workaround — the structure is the subject

Free and local tools help individuals, but they don't undo the structure: a few firms own the frontier, the labor that trains and moderates these systems is hidden and underpaid, and automated decisions land hardest on people with the least room to contest them. The camp's Further Reading carries the texts that name this directly — read them as the other half of the access question.

Atlas of AI — Kate CrawfordAI as an extractive industry: lithium, labor, data, power Automating Inequality — Virginia EubanksAutomated decision systems aimed at poor families Ghost Work — Mary L. Gray & Siddharth SuriThe hidden human labor inside every "automated" system Resisting AI — Dan McQuillanRefusal and restructuring as legitimate positions, not failures to adapt

The take-home: no one has to pay to do this camp, and "I can't afford the tools" is never a reason to feel shut out of understanding them. Access is a structural problem worth naming out loud — and the same investigation that makes a model's defaults visible can make the cost of using it visible too.