Perceptions of AI Integration in Non-STEM Education: Exploring Practical Implementation, Ethical Acceptance, and Educator Collaboration in Real-World Classrooms

Artificial Intelligence (AI) is revolutionising educational practices, yet its
integration has predominantly concentrated on STEM disciplines, leaving non-STEM fields like
the humanities, social sciences, and arts relatively underexplored. Recognising this imbalance, the
current study aims to investigate the perceptions, practical applications, ethical concerns, and
collaborative opportunities surrounding AI integration in non-STEM education. This research
shifts the focus toward real-world classroom experiences, offering a comprehensive examination
of how AI technologies are transforming teaching and learning in diverse academic fields outside
the traditional STEM areas.
The literature review highlights that although AI tools like Intelligent Tutoring Systems,
language learning apps, and creative AI have made their way into non-STEM subjects, their
adoption remains fragmented. A major concern is that many AI systems in non-STEM fields are
adapted from STEM-based models, often failing to accommodate the subjective, interpretive, and
creative nature of humanities education. Additionally, the literature reveals a consistent lack of
educator involvement during the design and deployment of AI tools, which exacerbates challenges
in alignment with pedagogical goals.
A significant research gap exists not only in the limited focus on non-STEM applications
but also in the scarcity of empirical, field-based studies that observe AI’s practical impact in
authentic classroom settings. Ethical issues—including data privacy, algorithmic bias, and the
potential erosion of creative expression—remain largely understudied in the context of non-STEM
education. This study seeks to bridge these gaps by grounding its investigation in real classroom
environments while giving voice to educators' firsthand experiences and concerns.
Utilising a mixed-methods approach, the study collected data through surveys, semi
structured interviews, and classroom observations involving non-STEM educators. Quantitative
analysis revealed that although AI tools were moderately used for administrative efficiency and
personalized content delivery, their pedagogical impact was limited without deeper customization.
Qualitative findings underscored widespread concerns about ethical transparency, bias in AI
feedback mechanisms, and a perceived disconnect between AI developers and educators' real
needs.
The findings further illuminate that most educators had minimal involvement in AI tool
development or selection processes, resulting in dissatisfaction with the rigidity and irrelevance of
many AI systems for non-STEM subjects. Ethical apprehensions were particularly heightened
among experienced teachers, who expressed concerns about how AI could undermine critical
thinking, creativity, and equitable access to learning resources. Importantly, teachers advocated for
more participatory approaches where they could collaborate actively with AI designers.