evidoria

← Back to browse

Good practice Imported

Innowacje Technologiczne w Szkole — Google for Education AI Pilot in Two Polish Schools

Poland · Siedlce · See the Poland profile

Evidence: Quasi-experimental Top 89% 27/100 · Ask Evidence Copilot about this practice

A five-month Google for Education/IBE-PIB pilot trained grade-6 teachers in Siedlce and Głoskowo, Poland to use AI for feedback; engagement rose but the government evaluator found AI-assisted students scored 17% lower on comprehension tests.

15 %
Full engagement rate before pilot
37 %
Full engagement rate after pilot
83 %
Teachers reporting faster feedback
-17 %
Comprehension/concept test score gap, AI-assisted vs comparison group
~75 participants
Participants at the Siedlce school
Innowacje Technologiczne w Szkole — Google for Education AI Pilot in Two Polish Schools Innowacje Technologiczne w Szkole — Google for Education AI Pilot in Two Polish Schools

Details

Maturity
Pilot
Promoter
Google for Education Poland / IBE-PIB (Instytut Badań Edukacyjnych)
Period
April–September 2025
Keywords
primary education, teacher training, edtech, government research institute

Context

Google for Education Poland, working with government research institute IBE-PIB (Instytut Badan Edukacyjnych) and hardware partners Intel, Acer and GoCloud, ran a five-month pilot from April to September 2025 testing AI tools and Chromebooks with grade-6 teachers and students at two schools, described by Google as the first programme of its kind in Poland.

Activities

The pilot studied AI-driven feedback in the classroom rather than just device deployment, and included a teacher AI-competency certification track.

Results

Self-reported student 'full engagement' rose from 15% to 37% between the pre- and post-pilot surveys, and 83% of participating teachers said AI let them give students faster feedback. IBE-PIB's evaluation also reported that students who used AI assistance scored 17% lower on comprehension and concept tests than the comparison group, attributed to 'cognitive offloading' -- reduced independent cognitive effort when AI does part of the thinking.

Conclusions

The pilot involved about 75 participants at one school alone, ran across only two schools, and no public evidence of continued funding or wider rollout beyond the five-month pilot period has been found -- a genuine negative finding reported honestly rather than spun positive.

Implementation
Indicative cost
Low (< €50k)
Time to results
Short (< 1 year)
Staffing & skills
Google for Education Poland, IBE-PIB (Instytut Badan Edukacyjnych) -- government research institute, evaluator, Intel, Acer and GoCloud (hardware partners)

Conditions for success

  • Independent government-institute evaluation design with a comparison group
  • Teacher AI-competency certification track

Common failure modes

  • Cognitive-offloading effect: AI-assisted students scored 17% lower on comprehension/concept tests than the comparison group
  • No evidence of continued funding or wider rollout after the five-month pilot

Commonly funded by

Own resources / municipal budget National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Replication kit

Reusable artefacts from this practice — as published by their sources.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful