RAND survey: higher-income parents use AI more, and their teens now use it more for homework
What happened
On 12 November 2026 RAND published a research report titled “Higher-Income Parents Use AI More and Encourage Their Children to Do the Same,” by Heather L. Schwartz and Melissa Kay Diliberti, sponsored by the Gates Foundation. It draws on two RAND survey panelssurvey panel A standing group of people who have agreed to answer surveys repeatedly, so researchers can ask them new questions over time., the American Parents Panel and the American Youth Panel. Parents were surveyed in February 2025 and May 2026, and teens aged 12 to 18 in May 2025 and May 2026, with roughly 1,000 parents and 1,800 teens. A subset of 533 teens could be matched to 427 of their parents.
The headline numbers:
- Parents who use AI for work or personal reasons rose from 54 percent in 2025 to 72 percent in 2026. In 2026 the rate was 84 percent in families earning $150,000 or more and 63 percent in families earning under $50,000. The gap between the highest- and lowest-income groups grew from 15 to 21 percentage pointspercentage point The plain difference between two percentages, so a move from 40 to 45 percent is a rise of 5 percentage points..
- Among parents, 67 percent in the highest-income group encouraged their children to use AI for schoolwork, against 45 percent in the lowest. Seventeen percent of the highest-income parents encouraged it wherever school rules allowed.
- Teens’ homework use was 39 percent in both income groups in spring 2025. By spring 2026 it was 57 percent among teens from families earning $150,000 or more and 43 percent among those under $50,000.
- In the matched sample, 81 percent of teens who perceived parental encouragement used AI for homework, against 38 percent who perceived discouragement.
What is claimed vs. what is verified
The percentages above are RAND’s own survey results; no independent group has re-analysed the data. The independent coverage found for this post, from the research-administration body CASRAI, restates RAND’s figures accurately, adds its own commentary, and points out that the data are self-reported. It does not test them.
Some statements go further than description. RAND says the association between perceived parental encouragement and teens’ use was stronger than any other factor it examined except age, including family income. That is the authors’ analysis of a survey, so it describes what goes together, not what causes what. Lead author Schwartz is quoted in RAND’s press release saying parents have “as vital a role as schools” in helping teens with AI; that is an interpretation.
At least one widely circulated newsletter summary described the 15-to-21-point gap as a gap in teens’ schoolwork use. In the report it is a gap in parents’ own use. The teen figures show a different pattern: no gap in 2025, a 14-point gap in 2026. The report pages read for this post do not state margins of error for that change.
Why it matters
The finding bears on a long-running question about new technology in education: whether it narrows or widens differences between richer and poorer students. A year ago these teens used AI for homework at the same rate regardless of family income; now they do not, and the change tracks parents’ own use and attitudes. The report itself says it is uncertain whether more AI use among higher-income teens will widen or narrow achievement gaps, because that depends on how the tools are used.
The authors also make recommendations: that states and districts give schools clear guidance for families, consider blocking general chatbots on school devices during students’ first attempts at assignments, and that developers build guardrails rather than outright blocks. These are the authors’ proposals, not results of the survey.
Open questions and other views
Researchers who see AI tutoring as a way to give every student individual help may read the widening gap as an access and awareness problem that schools could close. Others, including teachers worried about students skipping the effortful part of learning, may read higher use as a mixed signal, since homework done by a chatbot is not the same as homework done with one. The survey asks whether teens use AI, not how, so it cannot settle which reading fits.
There are also measurement questions. Self-reports of AI use can shift with social expectations, and parents who use AI may be more likely to notice or report their children’s use. Matching 533 teens to their parents allows a direct comparison of perceptions and behaviour, but it is a subset of the larger panel, and how representative it is matters for the 81-versus-38 figure. Finally, two time points show a change but not a trend; another year of data would show whether the teen gap keeps growing, stabilises or closes.
Sources
- RAND Corporation, Schwartz, H. L., and Diliberti, M. K., “Higher-Income Parents Use AI More and Encourage Their Children to Do the Same,” research report RR-A4742-5, 12 November 2026. Link
- RAND Corporation, press release, “Higher-Income Parents More Likely to Use AI, Encourage Teens to Do the Same,” 12 November 2026. Link
- CASRAI Editorial Board, “RAND: Teen AI Homework Use Splits by Income,” 15 November 2026 (independent summary and commentary). Link