2026 Edition · Published 10 February 2026

AI Fluency in K-12: A Seven-Country Teacher Baseline

How 4,800 school teachers across India, the United States, the UAE, the United Kingdom, Singapore, Saudi Arabia and Australia actually use AI, and what their schools have written down about it.

n = 4,800 teachers · 7 countries · fieldwork Oct 2025 to Jan 2026

Key finding

In the NASCA seven-country baseline of 4,800 K-12 teachers, 71 percent use a generative AI tool at least weekly, while only 18 percent report a formal school policy conversation about it. The gap between everyday classroom use and written school guidance is the single largest finding of the 2026 survey.

Fieldwork ran from October 2025 to January 2026 across India, the United States, the UAE, the United Kingdom, Singapore, Saudi Arabia and Australia. The full tabulation, methodology and CSV are published free on this page.

Headline numbers

71%
of teachers use a generative AI tool weekly or more oftenHighest in the UAE and Singapore, lowest in the United Kingdom.
18%
say their school has held a formal conversation about AI policyWritten, minuted guidance rather than an informal staffroom view.
3.4x
more teachers use AI than have been trained on itOnly 21 percent report any structured AI training in the last year.
62%
want assessment redesigned rather than AI bannedPreference is consistent across all seven countries.

The data

Table 1. Weekly AI use, formal policy and training, by countryPercentage of surveyed teachers. Rounded to the nearest whole number.
CountryTeachers surveyedWeekly AI use (%)Formal policy conversation (%)Structured training in last year (%)
India1420741619
United States980692427
United Arab Emirates640812226
United Kingdom560611922
Singapore470792125
Saudi Arabia420721417
Australia310661820
All countries4800711821
Table 2. What teachers use AI forMultiple selection permitted, so columns do not total 100 percent.
Use caseShare of AI-using teachers (%)
Lesson planning and resource drafting68
Differentiating work for mixed ability44
Writing feedback and report comments41
Building quizzes and assessment items37
Explaining a concept to themselves first33
Administrative email and parent communication29
Co-teaching with students in the room12
Table 3. What teachers say they need next
Stated needShare of all teachers (%)
Assessment designs that survive AI62
A clear written school policy57
Hands-on training with their own subject54
Age-appropriate guidance for students48
Time in the timetable to plan with AI39
An outright ban on AI in school8

Every table on this page is available as a single CSV file: download the dataset.

What the data shows

Use is ahead of governance in every country surveyed

No country in the sample showed policy conversation rates within 40 points of weekly use. Teachers are making individual judgement calls daily, without a written position from their school to lean on.

Training is the constraint, not willingness

Teachers who received any structured AI training in the last year reported higher confidence in every category, including talking to parents about AI. Only about one in five had that training.

Teachers want assessment redesigned, not AI banned

Support for an outright ban sat at 8 percent. The dominant request was assessment that rewards process, reasoning and defence of work, which is what an AI-resilient rubric is built to do.

Student-facing use is still rare

Only 12 percent of AI-using teachers had used AI alongside students in the room. Most use is invisible preparation, which means students see the output but never learn the method.

Methodology

  • Population: practising K-12 teachers in India, the United States, the United Arab Emirates, the United Kingdom, Singapore, Saudi Arabia and Australia.
  • Sampling: stratified by country and school stage, drawn from NASCA partner schools, World STEM Federation member networks and open recruitment through teacher associations.
  • Instrument: a 24-question online survey, English medium, median completion time 9 minutes.
  • Fieldwork window: 6 October 2025 to 24 January 2026.
  • Cleaning: responses under 3 minutes, duplicate device fingerprints and incomplete country blocks were removed before tabulation.
  • Weighting: country results are unweighted; the all-country row is the simple pooled percentage across the cleaned sample.

Limitations

  • Recruitment leans towards schools already engaged with STEAM programmes, so absolute AI use is likely a little higher than the global teacher average.
  • The survey was English medium only, which under-represents teachers in non-English medium schools.
  • Self-reported behaviour is not observed behaviour. Frequency claims are directional.

Questions this report answers

What percentage of teachers use AI in the classroom?

In the NASCA 2026 seven-country baseline of 4,800 K-12 teachers, 71 percent use a generative AI tool at least weekly, mostly for lesson planning, differentiation and feedback. Only 12 percent of those teachers have used AI alongside students in the room.

How many schools have a written AI policy?

Only 18 percent of teachers in the NASCA 2026 baseline report a formal, minuted conversation about AI policy at their school, against 71 percent weekly use by teachers themselves.

Do teachers want AI banned in schools?

No. Support for an outright ban was 8 percent. 62 percent asked instead for assessment to be redesigned so that it rewards reasoning and defence of work.

Can I cite or reuse this data?

Yes. The dataset is published under CC BY 4.0. Cite it as: NASCA Research Desk with the World STEM Federation (2026), AI Fluency in K-12: A Seven-Country Teacher Baseline, https://www.nasca.edu.in/research/reports/ai-fluency-baseline-2026.

How to cite

NASCA Research Desk with the World STEM Federation (2026). AI Fluency in K-12: A Seven-Country Teacher Baseline. NASCA. https://www.nasca.edu.in/research/reports/ai-fluency-baseline-2026

Licensed CC BY 4.0. Quote the figures freely, with attribution and a link back to this page. Media and researchers may request the raw tabulation and the survey instrument.

Want the raw tabulation?

We share the instrument, the cleaned dataset and an interview with the research desk on request.