[ GUIDE ]

Equal access to strong AI: what schools can promise

Paying for AI tracks income, and blocks fall hardest on students without their own devices. What a school can honestly promise about equal access, and what it should measure.

Plan a pilot Watch the demo

Facts as of Oct 18, 2026Last reviewed Oct 19, 20263 min read

1 in 5

teens in households under $30,000 do all or most schoolwork with chatbot help, against 7% in higher-income homes

Source: Pew Research Center, Feb 2026

[ EQUITY ]

The short answer

A school can give every student the same strong AI for schoolwork, and that matters: paying for AI tracks income, and blocks fall hardest on students without their own devices. But access alone rarely closes learning gaps. The promise is equal access under learning rules; whether it changes learning is something to measure.

Who pays for AI

About 3% of Bank of America households paid for AI services in February 2026, and households earning over $125,000 were the largest group of paying households. 66% of upper-income US adults have used AI chatbots, against 41% of lower-income adults (Pew Research Center, 2026). In a 2024 Harvard survey, 40% of undergraduate AI users without financial aid paid for AI products, against 20% of those with partial or full aid; that is one college. No survey measures how many teens pay.

Lower-income teens use AI more, mostly free

The divide is not AI or no AI. About 1 in 5 teens in households earning under $30,000 do all or most of their schoolwork with chatbot help, against 7% in higher-income households (Pew Research Center, Feb 2026). The difference is which AI, on which device, used how.

Free and paid AI are different products

Free ChatGPT runs on OpenAI's smallest Luna models; the $20-a-month Plus plan adds GPT-6 Sol reasoning. Stronger models make fewer errors: in a study of middle-school math explanations, GPT-4's were valid 94% of the time, against 10-43% for GPT-3.5. A student on a free tier and a student whose family pays are not using the same tool.

Blocks fall on students without devices

78% of teens in households under $30,000 have a home desktop or laptop, against 93% in households at $75,000 or more. In Los Angeles, students with their own devices keep using AI at home, while students who rely on district devices cannot.

Access alone rarely closes gaps

Programs that only give students computers or internet access have generally not raised K-12 test scores. Training is uneven too: in fall 2024, 67% of low-poverty districts had trained teachers on AI, against 39% of high-poverty districts. And a strong model without learning rules can make outsourcing easier: answer-giving AI cut unaided exam scores 17% in a field experiment.

Students with IEPs and 504 plans

Teens with an IEP or 504 plan tailor AI tools to their learning needs more often than other teens: 44% against 34% (Common Sense Media, Aug 2026). Common Sense warns that blanket bans can cut off tools these students need. Read AI for students with IEPs and 504 plans.

What makes the access promise real

The same models for every student in an approved class, whatever their family pays at home.
A fair share for each student: per-student daily limits, sized so that heavy use by a few does not crowd out everyone else.
Learning rules on every class chat, so stronger AI does not mean easier outsourcing.
Teacher training before student access.
Equity measures in the pilot without collecting family income: device reliance, home internet and paid-AI access, reported in groups of 10 or more.

[ FAQ ]

Questions and answers

Does school AI narrow achievement gaps?

No study shows that. A school AI can equalize access, but access alone rarely raises test scores. Measure learning in a pilot.

Do lower-income students use AI less?

Not for schoolwork. About 1 in 5 teens in households under $30,000 do all or most schoolwork with chatbot help, against 7% in higher-income homes (Pew, Feb 2026).