Speak Up report finds students learning AI outside class outpace classroom learners

Project Tomorrow released its 23rd annual Speak Up report on October 28, bringing student experiences with artificial intelligence into a wider discussion about classroom technology. The nonprofit's public briefing findings identify a gap between high school students' independent exploration of AI and their experience learning about it in class. That is a finding about reported exposure, rather than a measurement of what students have mastered.

Among high school respondents, 46% reported learning about AI through independent experimentation, compared with 30% who reported learning about it in class. Those categories should not be treated as mutually exclusive groups. A student may encounter the technology in both settings. The comparison also does not establish that experimenting alone produces better understanding than classroom instruction.

The nonprofit's dated release describes input from more than 95,000 K-12 students, educators, administrators and families across rural, urban and suburban schools. That total covers multiple stakeholder groups; it is not a sample of 95,000 high school students. Project Tomorrow says the research is supported by education-sector sponsors, including ClassDojo for Districts, Edmentum, edWeb.net and PowerSchool, alongside Agile Education Marketing.

Government Technology's independent coverage reports on both the findings and the accompanying student panel. It describes a Virginia seventh grader's classroom practice of tracing AI responses to sources and checking reliability. A California eighth grader described comparing generated math solutions with her own answers. These are accounts of particular learning practices, rather than controlled evidence of an improvement in grades.

The publication also reports that daily or weekly AI use for studying and research rose from 23% of high school respondents in 2023-24 to 40% in 2025-26. Only 11% of teachers reported feeling very comfortable allowing students to use AI resources. These figures describe different questions and respondent groups, so they cannot be combined into a single measure of school readiness.

The practical implication is to examine what students do with an answer: whether they check its origin, test its reasoning and retain responsibility for their work. Project Tomorrow advocates purposeful digital instruction and durable skills. Its report and the panel provide perspectives for that debate, while leaving the causal question open. The public summaries do not establish national representativeness or prove that a particular AI tool improves learning. Decisions about classroom use still require evidence about the specific task, students and teaching approach.