Beyond the Hype: How Higher Education Stakeholders View the Benefits and Concerns of Generative AI for Teaching, Research, and Administration

Synthesis: Humble and Mozelius (2026) conduct a SWOT analysis of 167 contributions collected via Padlet during webinars at two Swedish universities attended by 152 higher-education personnel. Strengths include content creation, translation, and transcription capabilities that streamline workflows and support collaboration; weaknesses include inaccuracies, lack of system transparency, and user dependency; opportunities include supportive roles, outsourcing repetitive tasks, and partnering with AI for innovative teaching; threats include destabilization of educational norms, ethical issues around data privacy and copyright, and organizational challenges. The study translates findings into a Balanced Scorecard framework across student, internal-business, innovation-and-learning, and financial perspectives.

Key Findings

• Stakeholders identified strengths in GAI content creation, translation, and transcription that streamline workflows and support collaboration.
• Weaknesses centered on inaccuracies (hallucinations, made-up references), lack of system transparency ('black box'), and user dependency.
• Opportunities included supportive roles, outsourcing repetitive tasks, and partnering with AI for innovative teaching practices.
• Threats included destabilization of educational norms, ethical issues around data privacy and copyright, and organizational challenges such as workforce adaptation.
• Stakeholders requested clear institutional guidelines, targeted professional development, and allocation of time and resources for responsible adoption.

What this means for practice

• Administrators. Convert the SWOT themes into a Balanced Scorecard with student, internal-business, innovation-and-learning, and financial perspectives, rather than debating strengths and threats as separate topics.
• Administrators. Fund the three things stakeholders actually asked for — clear institutional guidelines, targeted professional development, and dedicated time and resources — instead of running ad hoc tool rollouts.
• Instructors. Pair GenAI adoption with explicit verification routines for the weaknesses stakeholders named most often: hallucinations and made-up references, opaque "black box" behavior, and user dependency.
• Instructors. Teach around the organizational threats as well as the tool: workforce adaptation and destabilized educational norms were cited alongside data privacy and copyright, so software training alone is insufficient.
• Researchers. Reuse the SWOT-to-Balanced-Scorecard mapping as a template for multi-stakeholder perception studies, since it bridges what personnel observe about GenAI with actionable institutional strategy.

Limitations

• The analysis rests on 167 Padlet contributions from 152 higher-education personnel at two Swedish universities, and the authors state the findings are context-specific insights rather than a globally representative picture.
• Fully anonymous Padlet collection prevented follow-up questions and made stakeholder roles unidentifiable, so the study could not test whether teachers, researchers, and administrators hold different views, as planned.
• Padlet participation was not mandatory and attendance dropped part way through the webinars, so staff who left or declined to post are absent from the data.
• The Balanced Scorecard is conceptual and built from qualitative SWOT themes rather than validated indicators; reliability coefficients, subgroup contrasts, and content-validity checks were not feasible with anonymous qualitative data.

Citation

Humble, N., & Mozelius, P. (2026). Beyond the hype – How higher education stakeholders view the benefits and concerns of generative AI for teaching, research, and administration. Computers and Education Open, 10, 100381.