K-12 Teacher AI Evidence Arc

From Adoption to Institutional Response

Three weighted national teacher snapshots show a changing AI landscape: instructional adoption in 2024, expected teaching impacts in 2025, and institutional guidance in 2026.

These are not repeated measures from one longitudinal survey. The sequence is an evidence arc, not a trend line.

Read the evidence

A three-wave evidence arc

Select a view. Every value remains visible without hovering.

2024 Adoption
2025 Expected teaching impact
2026 Institutional response

The central asymmetry

Fall 2025 teachers expected teachers' and students' AI use to affect teaching differently.

Teachers' AI use

72.5%

expected teachers' AI use to make teaching easier.

Same survey wave.
Different valid item denominators.

These items concern the job of teaching; they do not measure student learning.

Students' AI use

61.9%

expected students' AI use to make teaching harder.

Interpretive claim

Taken together, the surveys describe an implementation problem: reported teacher use was widespread by Fall 2025, respondents expected teachers' and students' AI use to affect teaching differently, and clear, formal, task-specific guidance remained limited. The evidence supports inquiry into infrastructural conditions and educator sensemaking. It does not directly measure pedagogical friction, governance quality, or student learning outcomes.

Where this belongs in the dissertation

Use the same evidence differently before and after primary data collection.

Chapters 1-3

National structural context

Use now to establish the problem context, justify the focus on institutional conditions, and document the secondary-data analysis plan.

  • Chapter 1: concise empirical context for the problem.
  • Chapter 2: connect the adoption-governance tension to the literature.
  • Chapter 3: identify datasets, weights, variables, limits, and integration plan.

Chapters 4-5

Later comparison frame

Reuse selected indicators only after primary analysis, when they can sit beside bounded-case themes in a joint display.

  • Do not treat national percentages as confirmation of participant accounts.
  • Compare convergence, divergence, and silence across evidence streams.
  • Keep meta-inferences qualitative-dominant and context-bounded.

Methods and limits

The visual is a descriptive synthesis of public-use survey files.

Analytic treatment

  • Weighted percentages use each public file's PORTAL_WEIGHT.
  • Missing and legitimate-skip responses are excluded from valid denominators.
  • Subgroup differences are descriptive; the visual makes no significance claims.
  • No raw microdata or respondent-level records are included in this public repository.

Comparability boundary

AIRS 2024 and the Gallup-sponsored 2025 and 2026 ATP surveys differ in wording, timing, analytic population, and denominators. Percentages across waves should not be subtracted as though they measured identical constructs. The equity view draws on two additional context datasets — the Spring 2025 ATP resource survey and the December 2024 NCES School Pulse Panel — that measure different constructs and are read only as convergent directional signals.

Survey waves used in the evidence arc
Wave Source Public analytic cases Primary contribution
Spring 2024 RAND AIRS teacher survey 9,126 Instructional AI adoption, planning, and curriculum modification
December 2024 NCES School Pulse Panel ~4,000 schools Context: school AI policy and instruction by neighborhood poverty (equity view)
Spring 2025 Gallup-sponsored RAND ATP 2,167 Context: AI-tool resource adequacy by school composition (equity view)
Fall 2025 Gallup-sponsored RAND ATP 2,012 Adoption, expected teaching impacts, support, and policy clarity
Winter 2026 Gallup-sponsored RAND ATP 2,069 Policy presence, policy clarity, guidance, and institutional stance

Sources

Primary reports, public survey documentation, and published Gallup benchmarks.

  1. Kaufman, J. H., Woo, A., Eagan, J., Lee, S., & Kassan, E. B. (2025). Uneven adoption of artificial intelligence tools among U.S. teachers and principals in the 2023-2024 school year. RAND Corporation. Report page
  2. RAND Corporation. (2024). 2024 American Instructional Resources Survey: Technical documentation and survey results. Technical report
  3. Ash, A. M. (2025, June 25). Three in 10 teachers use AI weekly, saving six weeks a year. Gallup. Separate Spring 2025 contextual benchmark
  4. Ash, A. M., & Senseman, K. (2026, May 27). Most teachers receive no formal guidance on AI use. Gallup. Published benchmark
  5. Gallup. (2026). Walton Family Foundation-Gallup K-12 teacher research. Research hub
  6. U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics. (2024). School Pulse Panel 2024-25, December 2024 technology and digital literacy data release. Survey results
  7. RAND Corporation. (n.d.). RAND Survey Panels Data Portal. Fall 2025 and Winter 2026 public-use files

Values displayed here were independently calculated from the public portal files and may differ from published Gallup or RAND estimates that use non-public weights. Review the public proposal audit trail and 21-claim crosswalk before reusing an estimate.