On this page

Synthesis: Chen and eight co-authors turn the Teacher Panel they convened at the International Conference on AI and Pedagogical Innovations in Hong Kong in May 2026 into an argument about what faculty development should become. The panelists hold posts in educational development, provost's offices, education faculties, and a foundation, across nine institutions in eight countries, and the commentary is explicit that it synthesizes that conversation rather than reporting a study. Its four insights run in order of pressure: AI use is already embedded in student work, which makes assessment the first system to break; AI literacy is necessary but insufficient without disciplinary and ethical capability; universities must connect learning to the world beyond campus while renewing the campus as a site of experimentation and futures literacy; and institutions need the capacity to govern uncertainty rather than to adopt tools quickly. From those insights the authors derive seven roles for university teachers, from human-centered educator and disciplinary expert through learning architect, assessment (re)designer, technically competent professional, ethical steward, and educational experimenter. Their distinctive move is to insist that the roles need not sit in one person: they can be distributed across a teaching team, with the first two treated as universal foundations.

No primary data: this is a brief communication that reports the authors' synthesis of a panel they organized and moderated, with no instrument, sample, or coding procedure; every claim below is an argued position rather than a measured finding.

Key Findings

  • Assessment is where the pressure lands first, and the problem is evidentiary. The panel treats take-home essays, reports, summaries, problem sets, and generic written reflections as formats that "may no longer reliably reflect individual understanding, process, judgement, or capability." The question it keeps returning to is not whether a student cheated but what a polished artifact demonstrates when parts of it may have been generated, revised, translated, or enhanced by a machine. Assessment therefore has to produce evidence of process: oral explanation, staged development, process documentation, critical commentary, in-class performance, workplace tasks, and the ability to critique and improve AI output.
  • AI literacy is necessary and insufficient. The authors argue for moving past generic tool awareness toward disciplinary, professional, interdisciplinary, and transdisciplinary capability. Students need to understand how AI changes evidence, standards, judgment, practice, and responsibility inside a particular knowledge community, and they need the human capacities that AI makes more valuable rather than less: resilience, adaptability, ethical judgment, communication, collaboration, and the willingness to question or resist a tool.
  • Seven roles, only two of them universal. The proposed set runs from human-centered educator and disciplinary and professional expert, which every university teacher should hold, through learning architect and intellectual guide, assessment (re)designer, technically competent and critically informed professional, ethical steward, and educational experimenter and evidence-based practitioner. The authors describe the last five as capacities that must be present somewhere in the teaching enterprise, not as a checklist each academic must complete.
  • The unit of faculty development is the team, not the individual. In large courses where many lecturers teach parallel classes and set the same assessments, one or two teachers can redesign the tasks and rubrics for the whole course while others take learning architecture, tool evaluation, or the systematic collection of evidence. The authors describe the resulting teacher as a conductor who coordinates expertise rather than a soloist who supplies all of it.
  • The campus is a living laboratory, not a legacy asset. The third insight asks universities to build short practice-focused modules with employers, practitioners, alumni, and community partners, and to treat the campus as the place where students meet complexity, deliberate about values, and imagine alternatives. Futures literacy, following Gáspár, is presented as a competence with its own structure rather than a slogan: ways of using the future, the skills those uses require, and the habit of noticing the assumptions that shape them.
  • Governance capacity is the institutional bottleneck. The authors argue that the capability that will distinguish universities is not speed of adoption but the capacity to govern continual technological uncertainty. That requires AI policy literacy: the ability to read policy, negotiate trade-offs between innovation and accountability, and participate in writing the rules rather than only implementing them.
  • Existing provision is uneven by the panel's own reading of the evidence. The commentary reports that faculty development gains are greatest in technical knowledge, confidence, and initial experimentation, and much weaker in pedagogical integration, ethical judgment, assessment policy, and understanding of what AI cannot do. Provision is also described as fragmented and technology-led, with comparatively little attention to inclusion and ethics.

How the study was done

There is no study in the empirical sense. The article is a brief communication: a structured argument built from a panel discussion at the International Conference on AI and Pedagogical Innovations, held at The Hong Kong Polytechnic University on 22 May 2026, with Julia Chen moderating. All nine authors were panelists, and the paper states plainly that the discussion was situated in a particular institutional and regional context even as the authors reached for international relevance.

The author list is itself the evidence base. It represents nine institutions across eight countries and four continents, with posts in educational development centers in Hong Kong, London, and Auckland, a provost's office in Beijing, education faculties in Bangkok, Växjö, and Piraeus, a foundation in Rabat, and a center for special projects in Moscow. The commentary draws on that spread of vantage points and on published literature, including UNESCO's guidance and AI competency framework for teachers, the Australian framework for AI in higher education, and recent studies of faculty AI competency and of training gaps in university teaching staff.

The method, insofar as the word applies, is synthesis and role construction. The authors take four pressure points from the panel, derive seven teacher roles, and then map each role to a faculty development focus, a set of activities, and an illustrative vignette in a single framework table. The framework is presented as preliminary and skeletal, intended to help institutions ask which roles their mission requires, which every teacher should develop, which can be distributed, and what structures would support them. The paper reports no sample, no instrument, no response rate, and no analysis procedure, and its data availability statement says only that all data analyzed are contained in the article. A generative AI tool was used for language assistance, with wording reviewed and arguments developed by the authors.

What this means for practice

  • Faculty developers: stop selling tools and start mapping roles. A workshop catalog organized by tool encourages adoption without capability. The framework's practical use is diagnostic: list the roles a program actually needs, mark which are covered by whom, and treat the gaps as the development plan rather than as individual deficiencies.
  • Teaching teams: distribute the roles and name a lead for each. In a large multi-section course, assessment design, learning architecture, tool evaluation, and evidence collection can sit with different people. The authors' condition is that everyone keeps baseline technical and ethical competence even when only one colleague leads that role, so distributed expertise does not become distributed ignorance.
  • Program leaders: redesign assessment before rewriting the policy. The paper's chain runs from evidentiary collapse to task redesign to rubric revision. Process documentation, staged submissions, oral defense, and tasks that require students to explain, critique, and justify their work do more for integrity than a stricter acceptable-use paragraph, and rubrics have to reward transparency and judgment or students will keep optimizing the polish of the product.
  • Administrators: budget for policy literacy, not only for licenses. If teachers are expected to shape governance, they need forums, time, and standing to do it. The paper's account of policy as iterative and distributed only works if the institution treats practitioner input as evidence rather than as consultation.
  • Institutions with limited resources: stage the work. The authors propose starting with the two universal roles plus minimum technical and ethical capability, then building team-level capacity in learning architecture and assessment redesign, then capacity for disciplined experimentation and evaluation. The sequence is offered as a prioritization logic, not a rigid ladder.

Limitations

  • The argument rests on a conversation among its authors. Nine colleagues who convened and staffed a panel synthesize their own discussion. There is no instrument, no participant group beyond the authors, and no analytic procedure, so agreement is produced by the format as much as observed.
  • The framework is untested. The authors call it skeletal and explicitly not a performance checklist, but no institution is reported as having implemented it, and there is no evidence about whether role distribution improves teaching, assessment, or student learning.
  • The panel is senior and institutional. Every voice belongs to an educational developer, professor, central administrator, or foundation executive. Students, early-career teachers, adjuncts, and disciplinary staff outside the panel's fields are absent, although the roles are proposed for them.
  • Geographic spread is asserted rather than compared. Nine institutions in eight countries supply the paper's claim to international relevance, but the commentary does not analyze how policy environments, resources, or academic cultures differ across them.
  • The claims about existing provision come from secondary sources. The finding that gains concentrate in technical knowledge and confidence is attributed to other studies rather than assembled systematically here, so it inherits whatever limits those studies carry.

Citation

Chen, J., Gordon, C., Kadili, A., Khlaisang, J., Milrad, M., Sampson, D., Sun, H., Vyskubina, A., & Wang, D. (2026). Future Faculty Development in the Age of AI: University Teachers’ Roles, Needs, and Institutional Support. Frontiers of Digital Education, 4(1), article 8.

Embed this page

Copy the code below to embed this page in a learning management system or other website. The embedded version shows only the page content — it hides the site header, navigation, and footer.