Date
AI Overview
Why it matters
The results educators get from large language models depend on how they ask them to help, making prompt engineering essential for tailoring AI responses to specific classroom needs.
Study at a glance
- Question: Does training secondary school science teachers in prompt engineering improve their perceptions of ChatGPT as an educational assistant?
- Setting and data: Public and private junior and senior high schools in Greece, 14 secondary natural sciences teachers.
- Method: Pre- and post-intervention survey with non-parametric comparison (Wilcoxon signed-rank test) and bootstrapping.
- AI tools: ChatGPT
Key findings
- Secondary science teachers significantly improved their median perception scores regarding ChatGPT as an educational assistant after receiving prompt engineering training (Z = -3.296, p < .001).
- Bootstrapping analysis showed an estimated mean difference between pre- and post-intervention scores ranging from 0.682 to 0.909, indicating a statistically significant change.
- Following the intervention, teachers showed greater confidence in what they could accomplish with language models and an increased appreciation for AI systems as educational assistants.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
Artificial Intelligence and especially Large Language Models (LLM), such as
ChatGPT has revolutionized the way educators work. The results we get from LLMs
depend on how we ask them to help us. The process and the technique behind an
effective input is called prompt engineering. The aim of this study is to
investigate whether science educators in secondary education improve their
attitude toward ChatGPT as a learning assistant after appropriate training in
prompt engineering. The results of the pilot study presented in this paper show
an improvement in the previously mentioned teachers perceptions.
Research Categories
What is the application?
Who is the user?
Who age?
Study design
Methods
AI tools
Publication Source
Publisher
arXiv
