Barriers to artificial intelligence adoption among instructors of IT-related disciplines in higher education: a survey-based study

The integration of generative artificial intelligence (AI) into higher education faces contradictory barriers. While the traditional “digital deficit” model posits that a lack of instructors' digital competencies is the primary obstacle, its applicability to IT-related disciplines remains questionable. This study investigates the key factors hindering AI adoption among IT instructors and empirically tests the “competence paradox” hypothesis, grounded in the Technology Acceptance Model (TAM), the Unified Theory of Acceptance and Use of Technology (UTAUT), and Technology Trust Theory. A quantitative multicenter survey ( N = 105) of IT instructors across three universities in Kazakhstan was conducted. The results revealed high self-assessed AI skills (M = 3.86) and strong perceived usefulness of AI (M = 4.19); however, correlation analysis found no relationship between competence levels and integration readiness ( ρ = 0.010, p = 0.917). Instead of internal skill gaps, the critical bottlenecks were external facilitating conditions, specifically the risks of academic dishonesty (M = 3.89), lack of licensed software (M = 3.78), and data privacy concerns (M = 3.41). These findings highlight a dual trust deficit—both technological (risk of hallucinations, M = 3.70) and institutional—culminating in an “institutional crisis of trust.” The study concludes that IT instructors are technically equipped for AI but abstain from full integration due to the absence of a secure and regulated ecosystem. Consequently, universities must shift their strategic focus from basic digital literacy training to robust infrastructural support and ethical policy development.