What the study found
GPT-4o, a generative AI model, showed some ability to represent aspects of the nature of science and the nature of AI-influenced science. The study also found that it had both strengths and weaknesses in distinguishing between the two.
Why the authors say this matters
The authors argue that educational technologists need to fine-tune GenAI so it can communicate its own influence in producing scientific knowledge. The study suggests this could support ethical and responsible use of GenAI in science education and help classroom instruction foster students’ epistemic learning outcomes in science.
What the researchers tested
The researchers examined whether and how GPT-4o could represent the nature of science and the nature of GenAI-influenced science. They drew on the Family Resemblance Approach, a framework that organizes science into categories such as aims and values, methods and methodological rules, knowledge, and practices, and interviewed GPT-4o about these categories.
What worked and what didn't
GPT-4o demonstrated some aspects of the nature of science and the nature of GenAI-influenced science. It also showed both strengths and weaknesses in making epistemic differentiation, meaning distinguishing between how science works and how AI-influenced science works.
What to keep in mind
The abstract does not provide detailed limits of the study beyond focusing on GPT-4o and the Family Resemblance Approach categories. It also does not specify how generalizable the findings are to other AI models or classroom settings.
Key points
- GPT-4o could represent some aspects of the nature of science.
- GPT-4o could also represent some aspects of AI-influenced science.
- The model showed both strengths and weaknesses in distinguishing the two.
- The authors link these findings to possible classroom instruction design.
- The study used the Family Resemblance Approach categories of science.
Disclosure
- Research title:
- GPT-4o showed partial ability to represent science and AI-influenced science
- Authors:
- Kason Ka Ching Cheung, Wenhao Zhang
- Institutions:
- Chinese University of Hong Kong, Education University of Hong Kong
- Publication date:
- 2026-02-11
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with scholarly research without reading academic papers — one filtered digest, every Friday.