The AI Revolution in Education: Will AI Replace or Assist Teachers in Higher Education?

The paper frames a pressing question in the context of rapid advances in generative AI (e.g., ChatGPT): will AI replace teachers or assist them in higher education? Citing predictions that robots could replace teachers and media discourse about AI-driven job displacement, the study aims to explore perceptions and experiences of students and teachers toward generative AI to determine whether AI could replace teachers or instead work alongside them.

The literature review traces AI in education from early computer-assisted instruction to intelligent tutoring systems (ITS) and current tools such as virtual assistants, chatbots, and immersive technologies. Key research themes include adaptive learning, expert systems/ITS, and AI’s role in educational processes. Benefits reported include personalized and adaptive learning, real-time feedback, and automation of administrative tasks, potentially freeing teacher time for higher-level work (e.g., curriculum, mentoring). Dialogue-based tutoring and NLP-driven tools can support knowledge co-creation, language learning, and provide objective feedback and monitoring of student progress. However, multiple works highlight AI’s limitations: lack of sentience and emotions, difficulty with values and social norms, and inferiority in social-emotional domains central to teaching (empathy, trust, motivation). AI-student interactions may not match the educational value of human relationships, and overreliance on AI/online systems can hinder social skill development. Other concerns include algorithm reliability, need for human oversight, bias/inequity, and limited holistic/visionary thinking. The reviewed scholarship generally positions AI as augmenting, not replacing, teachers. The paper synthesizes eight categories (26 aspects) of uniquely human teacher strengths that AI cannot replicate yet, including emotional/interpersonal skills (human connection, cultural sensitivity, resilience building, trust, SEL, role modeling), pedagogical skills (real-world context, curiosity, PD, debate and open-mindedness, conflict resolution, experiential learning; critical thinking and creativity; collaboration and teamwork; nurturing creativity and innovation; life skills), ethical/moral guidance, personalized support (behavior management; special needs support), community and civic engagement (parent communication; civic engagement), career/personal mentorship, physical education/sports coaching, and artistic education. A collaborative paradigm is emphasized in the literature: human-AI synergy can outperform either alone, with examples such as AI teaching assistants (e.g., Jill Watson), humanoid robots assisting lectures, social robots for remedial math, conversational AI in language education, and AI-enabled assessment ecosystems. This body of work underscores the potential for AI to automate routine tasks and provide data-driven insights while teachers provide emotional support, social context, ethical guidance, and adaptive pedagogy.

The study employed a survey design targeting students and teachers at Hong Kong universities. An online questionnaire included closed- and open-ended items on integration of generative AI (e.g., ChatGPT) in higher education, perceived risks, and impacts on teaching and learning. Participants were recruited via bulk email using convenience sampling; informed consent was obtained. The final sample comprised 384 undergraduate and postgraduate students and 144 teachers from various disciplines. Descriptive statistics were used for quantitative analysis; thematic analysis was applied to open-ended responses. Two pilot rounds (n≈20 total students and teachers) informed questionnaire refinement through feedback and researcher discussions.

Quantitative findings (selected items; higher scores = greater agreement): - Openness to integrating generative AI in learning: Students M=3.86, SD=1.008 vs Teachers M=3.61, SD=1.183; t=2.238, df=215.111, p=.026 (students more open). - Guidance for coursework as effectively as teachers: Students M=3.08, SD=1.142 vs Teachers M=2.67, SD=1.127; t=3.636, df=531, p<.001 (students more favorable). - AI improves academic performance: Students M=3.47, SD=.979 vs Teachers M=3.29, SD=1.115; t=1.792, df=507, p=.074 (trend; NS at .05). - AI helps become better writers: Students M=3.32, SD=1.162 vs Teachers M=3.01, SD=1.273; t=2.577, df=525, p=.010. - AI provides unique insights/perspectives: Students M=3.74, SD=1.076 vs Teachers M=3.47, SD=1.079; t=2.533, df=526, p=.012. - 24/7 availability valued: Students M=4.13, SD=.826 vs Teachers M=3.69, SD=1.068; t=4.483, df=216.752, p<.001. - Students would ask AI questions they won’t ask teachers: Teachers M=3.73, SD=.883 vs Students M=3.39, SD=1.094; t=-3.695, df=308.968, p<.001 (teachers perceive this more strongly). - AI harms development of generic/transferable skills: Teachers M=3.48, SD=1.238 vs Students M=3.10, SD=1.227; t=-3.087, df=523, p=.002 (teachers more concerned). - Teachers’ confidence detecting AI use in assignments: Teachers M=2.20, SD=1.125 vs Students M=2.54, SD=1.102; t=3.066, df=486, p=.002 (teachers less confident). - Belief AI will replace teachers: Students M=2.02, SD=.919 vs Teachers M=2.03, SD=.946; t=-.057, df=539, p=.955 (both groups generally disagree). - Willingness to pursue fully online AI-assisted degree: Students M=2.70, SD=1.270 vs Teachers M=2.84, SD=1.220; t=-1.140, df=514, p=.255 (low support). Overall: Both groups are open to AI integration; students are generally more positive about benefits. Neither group expects AI to replace teachers. Teachers show heightened concern about impacts on generic/holistic skills and lower confidence in AI-use detection. Qualitative themes: - Replacement vs irreplaceability: A minority foresee possible replacement if AI achieves human-like teaching; most argue teachers are irreplaceable due to human thinking, creativity, emotions, cultural sensitivity, and value transmission. - Working with teachers: Reported benefits include enhanced course planning/design (brainstorming, scenario generation, question creation), support for writing and research (structure, clarity, references/keywords), preparing students for AI-driven workplaces, time efficiency and cost reduction (admin, logistics, templates), and enabling personalized tutoring and immediate feedback. - Working against teachers: Risks include low AI literacy leading to misuse, ethical/integrity concerns (plagiarism, bias, privacy), overreliance reducing originality and critical thinking, undermining holistic competency development, factual errors, and potential erosion of trust. Calls were made for clear guidelines, training, and governance.

The findings address the research questions by showing broad openness to AI’s educational potential while rejecting the notion that AI will replace teachers in the foreseeable future. Quantitatively, students report greater enthusiasm for AI’s benefits (guidance, writing support, 24/7 access), whereas teachers express concerns about developmental outcomes (generic/transferable skills) and detection challenges. Qualitative insights underscore that human teachers provide emotional intelligence, cultural sensitivity, ethical guidance, trust-building, adaptive pedagogy, and community engagement—qualities central to holistic education and difficult for AI to replicate. The literature and data converge on a collaborative model: AI augments teachers by automating routine tasks, providing data-driven feedback, and enabling personalization, while teachers focus on higher-order teaching, mentorship, and socio-emotional development. Effective integration requires AI literacy for staff and students, ethical frameworks (privacy, bias, academic integrity), equitable access, and pedagogical redesign to leverage AI strengths without undermining critical and creative capacities.

The study concludes that while AI is increasingly capable and valued by both students and teachers, participants generally do not believe it will replace human educators. Instead, the future lies in synergistic human–AI collaboration. The paper offers a roadmap that highlights uniquely human teacher strengths across emotional, pedagogical, ethical, personalized support, community engagement, mentorship, physical education, and arts domains. Realizing this vision requires upskilling in AI literacy, thoughtful curriculum and assessment design that harness AI for efficiency and personalization, and governance to ensure ethical and equitable use. The work reinforces evidence that teachers remain the most influential factor in student learning, while cautioning that educators relying solely on traditional, content-driven methods risk obsolescence in an AI-enabled landscape. Future research should examine longitudinal impacts of AI-integrated pedagogies on holistic competencies, effective detection and integrity mechanisms, scalable AI literacy interventions, and comparative effectiveness of human–AI collaborative models across disciplines and contexts.

The study’s limitations include: (1) a comparatively small, convenience sample that may not represent all post-secondary institutions; (2) focus limited to text-based generative AI, excluding other AI modalities; and (3) reliance on self-reported data, which may introduce bias or inaccuracies.