evidoria

← Back to browse

Good practice Imported

INSIGHT -- A Pre-Post Evaluation of a 3-Hour Generative-AI Training for Teachers in Potsdam, Germany

Germany · Potsdam · See the Germany profile

Evidence: Observational / pre–post Top 81% 33/100 · Ask Evidence Copilot about this practice

University of Potsdam researchers ran a 3-hour AI training with 100 teachers across five Brandenburg schools, finding significant gains in perceived usefulness (p=.002) and ease of use (p=.015) -- but no significant change in intention to actually use the tools.

100 teachers
Teachers with matched pre-post responses analysed
p=.002, d=-0.32
Perceived Usefulness change (significant increase) (March-September 2023)
p=.253
Behavioural Intention to use AI tools change (not significant) (March-September 2023)

Details

Maturity
Pilot
Promoter
University of Potsdam -- Inclusive Education with Focus on Language
Period
March-September 2023 (published September 2026)
Region (NUTS)
DE40
Keywords
Teacher professional development, education research

Context

INSIGHT ("Information on Schooling with Generative and Helpful AI Tools") is a three-hour, in-person professional-development session developed by Janne Mesenhoeller and Katrin Boehme of the University of Potsdam to introduce teachers to large language models and their use for differentiated instruction, delivered between March and September 2023 across five Brandenburg schools (three primary, one combined primary-secondary, one academic secondary/gymnasium).

Objectives

The training aimed to shift teachers' perceived usefulness, perceived ease of use, and behavioural intention to use generative-AI tools for differentiated instruction, measured with the SAELKIS instrument.

Activities

Of 112 teachers who completed a pre-training questionnaire and 106 who completed the post-training version, 100 matched pre-post responses were analysed using paired-samples t-tests.

Results

Perceived Usefulness rose significantly (t(99) = -3.24, p = .002, d = -0.32) and Perceived Ease of Use rose significantly (t(99) = -2.47, p = .015, d = -0.25), but Behavioural Intention to use AI tools showed no significant change (t(99) = 1.15, p = .253).

Conclusions

Published in Frontiers in Education on 16 September 2026, the study is transparent about this limitation: a short, single-session, single-region training shifted teachers' attitudes toward AI but not yet their stated intention to adopt it, and its in-person format was not designed for scaling beyond the five participating schools.

Implementation
Indicative cost
Low (< €50k) — No budget figures published; a single 3-hour in-person training session per school.
Time to results
Short (< 1 year) — Delivered March-September 2023; published in Frontiers in Education on 16 September 2026.
Staffing & skills
Janne Mesenhoeller and Katrin Boehme, University of Potsdam, participating teachers across five Brandenburg schools

Conditions for success

  • a structured, validated measurement instrument (SAELKIS) with a matched pre-post design
  • peer review and transparent reporting of null results

Common failure modes

  • a single three-hour, in-person session with no reported scale-up plan
  • no significant change in teachers' behavioural intention to actually use AI tools
  • single-region (Brandenburg), single-session design limits generalisability

Where it fits

Governance type
university-led teacher professional-development research
Scale
five schools, one German state (Brandenburg)
Income level
high-income

Commonly funded by

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.

Similar practices you may find useful