# Girls Who Code survey links teen girls’ AI use with a sense of agency > A Girls Who Code survey finds higher reported AI use among teen girls who feel able to influence the technology. Its one-time, self-reported design cannot establish cause. By BIGCHANGE.AI Editorial Published: 2026-10-11T05:03:09.294Z Updated: 2026-10-11T05:03:09.294Z Canonical: https://bigchange.ai/blog/girls-who-code-teen-ai-agency-survey ![Conceptual charcoal illustration of a blank, anonymous survey form with empty response circles resting on a stack of paper.](https://bigchange.ai/api/media/file/girls-ai-agency-survey-hero-v1.png) An anonymous questionnaire illustrates a survey of teens' reported views and AI use; the survey does not establish that a sense of agency causes AI use. AI-generated conceptual illustration by BIG CHANGE. A [Girls Who Code survey of U.S. teens](https://assets.girlswhocode.com/craft-prod/images/GWC_AI_Ethics_Report_V7.pdf) finds that girls who believe they can influence AI report using it more often than girls who feel they have little say. The finding is useful for educators and parents, but it comes from a single survey at one point in time. It cannot show that building a sense of agency will increase anyone’s AI use. ## The big change - **What changed:** A national survey adds girls’ views about privacy and influence to a discussion often reduced to how frequently they use AI. It finds substantial use alongside clear concerns about the technology. - **Why it matters:** Schools setting AI rules and companies seeking teen users now have direct reports of what these respondents value. Girls in the survey wanted clearer guidance and ways to give feedback; a usage count alone would miss those requests. - **What to watch:** The practical question is whether institutions offer meaningful ways for students to shape rules and report problems. This survey identifies demand for those channels. It does not measure the effect of creating them. ## The usage gap depends on what is measured In the report’s general-population sample, 33% of girls and 47% of boys said they used AI tools daily. Confidence differed too: 42% of girls and 54% of boys rated themselves very confident using the tools (Figure 1, page 5). These are answers about behavior and confidence, not a measure of interest in AI. The same report says 75% of girls considered learning to use AI responsibly important for their future (page 6). Girls Who Code also grouped respondents by four statements about agency: whether they felt able to use AI in line with their values, influence people around them, affect AI companies through consumer choices, and see themselves involved in future AI work or decisions. Agreement with three or four statements counted as high agency; agreement with zero or one counted as low agency. The report does not place respondents who agreed with two statements in either extreme group (page 8). Among high-agency girls, 88% reported using AI daily or several times a week. That compares with 54% of low-agency girls, 73% of all girls and 83% of all boys (Figure 3, page 8). The comparison is between selected subgroups and broader groups, and it measures *frequent* use rather than the *daily* use in Figure 1. It should not be read as an experiment in which agency was changed and usage was observed afterward. The report also found that high-agency girls were more likely to say they had spoken up at school about AI use: 38%, versus 18% of low-agency girls. The base sizes for that action chart were 413 and 323 respectively (Figure 4, page 9). Those responses show an association between reported beliefs and reported action. The direction of any influence remains unknown; experience with AI, access, confidence and other differences could matter. ## Privacy and school rules appear in the answers In a question asking respondents to choose between protection and convenience, 85% of girls and 71% of boys chose keeping personal information private over receiving a more personalized answer (Figure 2, page 6). That choice says what respondents preferred in the survey scenario. It does not establish how a particular AI service handles their data or whether teens behave the same way when using one. School was the setting where surveyed girls said they felt most concerned about AI use. Yet 16% named school as the biggest influence on their views of AI, while 35% named social media. A separate question found 59% said social media was where they mainly learned about AI (pages 15–16). These questions measure different things: a source of learning, a source of influence, and a setting of concern. Taken together, they give schools a reason to ask students how rules and teaching relate to what they encounter elsewhere. The age comparison also needs care. Girls aged 17–18 reported less frequent use than girls aged 13–16, and the share saying they could not make a difference in how AI is built and used rose from 8% in the younger group to 14% in the older group (pages 13–14). This is a comparison of age groups surveyed at the same time, not evidence that an individual girl’s sense of influence declines as she grows older. ## What the sample can establish Girls Who Code built and administered the online questionnaire with Touchstone Research from April 22 to May 18, 2026. Its 2,328 respondents were U.S. teenagers aged 13–18. They comprised 1,858 teens recruited from national online consumer panels and 470 Girls Who Code alumni recruited by email. The panel group included 940 girls and respondents of other gender identities, plus 918 boys. The report says the sample was evenly distributed across age groups and nationally representative by region and race or ethnicity; it reports differences called “significant” at the 95% confidence level (methodology, page 20). Those details matter when interpreting the charts. The alumni are a separately recruited group with a direct connection to the organization, and the report’s “girls” panel category includes other gender identities. Base sizes vary by question. The published methodology does not specify weighting, response rates or a probability-based sampling procedure, so the percentages should be read as survey estimates with the stated design, not as a precise census of all U.S. teens. Girls Who Code produced the study and advocates for girls’ participation in technology. Its report is primary evidence for what these respondents said, while its recommendations for education and policy are the organization’s interpretation. Independent replication was not identified for this article. ### Sources & further reading - [Girls Who Code, *Agency in the Age of AI: A National Study on Girls and Ethical Technology*](https://assets.girlswhocode.com/craft-prod/images/GWC_AI_Ethics_Report_V7.pdf) (2026). The full report provides the charts, agency grouping and methodology. Its online survey records self-reported answers; it does not test whether agency changes AI use. - [Girls Who Code’s study announcement](https://girlswhocode.com/news/agency-in-the-age-of-ai-a-national-study-on-girls-and-ethical-technology). This summarizes the organization’s interpretation; the figures and methods above were checked against the full report. - [BIG CHANGE: Common Sense Media tests the limits of ChatGPT’s teen safety alerts](https://bigchange.ai/blog/common-sense-chatgpt-teens-safety-test). Related coverage of a specific product’s alert behavior, rather than a survey of teens’ views. ## Sources - [Agency in the Age of AI: A National Study on Girls and Ethical Technology](https://assets.girlswhocode.com/craft-prod/images/GWC_AI_Ethics_Report_V7.pdf) — Primary complete report, including figures and survey methodology; published by the organization behind the study. - [Agency in the Age of AI announcement](https://girlswhocode.com/news/agency-in-the-age-of-ai-a-national-study-on-girls-and-ethical-technology) — Organization press account; not independent corroboration of survey interpretation.