Perceptions and Acceptance of Artificial Intelligence in Science Education Programmes: Voices of Pre-Service Science Teachers

Synthesis: Amponsah et al. (2026) examine pre-service science teachers' perceptions, acceptance, and use of artificial intelligence in science education programs in Ghana, guided by the Unified Theory of Acceptance and Use of Technology (UTAUT) and the Theory of Planned Behavior. Drawing on questionnaire data from 380 pre-service science teachers, they find generally positive perceptions of AI — with participants recognizing its potential to improve teaching — favorable effort expectancy, and strong behavioral intentions to use AI in future lessons, yet only moderate actual classroom use. ChatGPT emerged as the most widely used tool (69% reported frequent use). The authors conclude that effective AI integration depends on closing the gap between favorable attitudes and actual practice, and on addressing teacher preparation.

Key Findings

1. Pre-service science teachers demonstrated positive attitudes, favorable effort expectancy, and strong behavioral intentions toward AI adoption, but actual classroom use remained moderate — an intention–use gap.
2. Performance expectancy (M = 3.80), attitude, and effort expectancy all scored significantly above the neutral midpoint on one-sample t-tests, indicating a meaningful positive trend.
3. ChatGPT was the most widely used AI tool, with 69% of respondents reporting frequent use; a small share reported never having used AI.
4. Social influence was comparatively weaker (M = 3.66), suggesting some hesitation in fully advocating AI use among peers.
5. The study used UTAUT and Theory of Planned Behavior as a combined lens, and reports reliability/validity metrics (Cronbach's alpha and composite reliability above 0.70).

Discussion

The study extends the technology adoption literature to the Ghanaian Science Education context, where teacher readiness and perceptions are decisive for successful educational-technology implementation. Its contribution is to foreground the gap between favorable attitudes and moderate actual use — a finding that complicates simple "acceptance" narratives and points to teacher competency and institutional support as the real levers for AI integration. For the knowledge base, it connects teacher education, AI Literacy, and technology-adoption theory, and reinforces the recurring theme that initial teacher enthusiasm must be matched by sustained professional development and facilitating conditions to translate into classroom practice.

What this means for practice

• Instructors. Convert positive attitudes into classroom use: assign small, concrete AI tasks (research, lesson planning, content explanation) that fit an existing lesson, since strong intentions alone left actual use moderate.
• Faculty developers. Extend training beyond tool basics to sophisticated pedagogical functions, where participants were least confident, using hands-on and sustained rather than one-off sessions.
• Faculty developers. Address facilitating conditions — digital infrastructure, curriculum alignment, and institutional support — because the intention–use gap was attributed to systemic barriers, not to attitudes.
• Administrators. Build peer norms around AI use: social influence scored lowest of the UTAUT constructs (M = 3.66), indicating that teachers hesitate to advocate AI use among colleagues.
• Instructors. Start from the tools teachers already use — ChatGPT was reported as frequently used by 69% of respondents — and scaffold from that familiarity toward pedagogical integration.

Limitations

• Reliance on self-reported data from 380 pre-service science teachers, which may be influenced by social desirability bias or overestimation of AI use and competence.
• Cross-sectional design captures perceptions and behaviors at a single point in time, limiting causal inference and tracking of adoption over time.
• Only teacher perspectives were collected; students and school administrators were not included, though they affect AI integration.
• Findings are specific to the Ghanaian context, and generalization to countries with different infrastructural and policy conditions should be treated with caution.

Citation

Amponsah, K. D., Adu-Boahen, J., Commey-Mintah, P., Kumassah, E. K., Ayittey, R. F., & Nketsiah, J. (2026). Perceptions and Acceptance of Artificial Intelligence in Science Education Programmes: Voices of Pre-Service Science Teachers. Education Quarterly Reviews, 9(3), 51–65. DOI: 10.31014/aior.1993.09.03.724.