On 12 November 2026, RAND published a survey report finding that artificial intelligence (AI) use among U.S. parents grew fastest in the highest-income families, and that an income gap has opened in how many teens use AI for homework. In spring 2026, 57% of teens in families earning $150,000 or more reported using AI for homework, against 43% in families earning less than $50,000. A year earlier both groups reported the same rate, 39%.
What RAND reported
The report, Higher-Income Parents Use AI More and Encourage Their Children to Do the Same (document RR-A4742-5), is by Heather L. Schwartz and Melissa Kay Diliberti and was sponsored by the Gates Foundation. RAND’s accompanying press release gives these figures:
- The share of parents using AI for work or personal reasons rose from 54% in 2025 to 72% in 2026. In families earning $150,000 or more it reached 84%, compared with 63% in families earning less than $50,000.
- Over 15 months, the gap in AI use between the highest- and lowest-income families widened from 15 to 21 percentage points, driven by surging use among the highest-income parents. This is a gap in parents’ AI use, not in teens’ use.
- Higher-income parents were more likely to encourage AI use for schoolwork: 67% versus 45% of lower-income parents. Seventeen percent of the highest-income parents said they encouraged their children to use AI wherever possible within school rules, though most parents favoured more limited uses such as brainstorming or checking work.
- Among teens whose parents frequently used AI, 63% reported using it for homework in at least some classes, compared with 40% of teens whose parents never used it.
The teen gap and the parent link
The income difference in teens’ own reported use is the newer finding: 57% versus 43% in spring 2026, from an identical 39% in both groups a year before. The report also examined 533 teens whose parents answered the same survey. Of those teens, 81% who perceived parental encouragement used AI for homework, compared with 38% who perceived discouragement. RAND states that teens’ perception of parental support correlated more strongly with their AI use than any other factor it examined, except age, and that this included family income, sex and parents’ perceptions of AI harms.
RAND’s authors say it remains uncertain whether greater AI use among higher-income teens will widen or narrow existing achievement gaps. The report’s topline, however, warns that unequal AI use potentially places the academic performance of teens in higher-income families at risk, and it cautions against unrestricted AI use on a student’s first attempts at schoolwork, which research suggests can undermine learning.
How the survey was done
RAND drew on four surveys fielded through the RAND American Parents Panel and the RAND American Youth Panel, both subsets of the RAND American Life Panel, a nationally representative survey panel. Parent surveys were fielded in February 2025 and May 2026. Teen surveys, of youth aged 12 to 18 enrolled in grades 5 to 12, were fielded in May 2025 and May 2026. The samples were about 1,000 parents and about 1,800 teens. A subset of 533 teens was matched to 427 parents in the same household, which allowed the within-family analysis above.
Two features of this design shape how far the results travel. The parent and teen figures come from self-report, and the family-level analysis rests on the matched subset rather than the full samples. Readers working with survey data of this kind can find background in our entries on survey research methods and sampling methods.
What it means for researchers and research offices
This is a primary-source finding about school-age families, so its direct relevance to universities is indirect, and the report does not study higher education. Three points are reasonable to draw from it, as interpretation rather than as RAND findings.
- Equity in AI access and guidance. Gaps in how early students are guided to use AI tools are a baseline that later institutional AI policies inherit. The report’s recommendations, aimed at states, districts, parents and AI developers, centre on explicit guidance and guardrails that separate AI doing the work from AI helping students refine their own thinking.
- AI literacy programmes. Institutions designing training can treat uneven prior exposure as a planning assumption. See our entry on AI literacy for the regulatory use of that term.
- Using the data. Panel surveys of this kind are a citable evidence base for education research proposals. The correct citation is to the RAND report and its DOI, with the parent-versus-teen distinction preserved.
Sources
- RAND, press release, Higher-Income Parents More Likely to Use AI, Encourage Teens to Do the Same, 12 November 2026.
- Schwartz, H. L. and Diliberti, M. K., Higher-Income Parents Use AI More and Encourage Their Children to Do the Same, RAND, RR-A4742-5, 12 November 2026, DOI 10.7249/RRA4742-5.
