AI for Good: Evaluating and Shaping Opportunities of AI in Education
By Isabel Fischer, Leda Mirbahai, and David Buxton
Following the rise of awareness of the opportunities (and threats) of artificial intelligence (AI) in education, we have created a task and finish group which aims to review and ‘imagine’ the opportunities and challenges of AI in education, incl. assessments. Our vision is to deploy AI as a tool to support all students, independent of background and socio-demographic characteristics, to be successful in their studies and in their future work, while ensuring academic integrity, as well as to support educators feel confident in using AI effectively in promoting learning. We are working in five (sub)groups:
- General AI in Education (AIEd) Opportunities & information sharing
- Novel and Diverse Assessment Designs
- Feedback, Marking, Authorship Detection
- Designing Teaching Content - ‘what is out there being developed?’
- 'Red Team': AI Ethics and Academic Integrity
As we are still interested in colleagues from within Warwick as well as other institutions and the wider community of stakeholders to join us, here some further information per (sub)group:
1) General AI in Education (AIEd) Opportunities & information sharing: We review how to capture, shape, and disseminate the opportunities for both learner-facing and educator-facing AI, mainly in HE but also considering how HE can support the secondary and even primary school sector (e.g. how to help teachers to experiment with different forms of AI in a low-stake environment). We also consider the benefits, such as reducing inequality, fairness and democratisation that AI offers, evaluating how we can support SDG 4 (equitable and quality education) and SDG 10 (reducing inequalities). We want to help educators to know how to potentially embrace recent AI developments for their professional practice. Combined with sub-group / Strand 5, the ‘red team’ we also want to inform colleagues on research (similar to mini literature reviews) on topics such as Algorithmic Fairness.
Target Output: A WIHEA page that is informative for colleagues new to AIEd (explanations, links to other resources, links to discussions / brainstorming exercises / blogs, suggestions for their practice)
2) Designing Assessments: We review the opportunities for designing and setting diverse assessments (Learner-facing), including embedding our work within our different Learning Circle’s work. It is in this strand that most of the student co-creation will take place.
Target Output: WIHEA page, blogs, and talks
3) Feedback and Marking: We review the opportunities of using AI for formative feedback (Learner-facing), summative feedback (Educator-facing), ‘AES – automated essay scoring’ (educator-facing), and stylometry (authorship authentication) as well as ChatGPT detection. One aspect of this strand (but not constrained to this strand) is also ‘Move fast, Break fast, Learn fast’ – doing small scale experiments and testing them (e.g., Consulting Students will experiment with mind maps this term and then can, but don’t have to, submit their work to the Warwick AI Essay Analyst for formative feedback and we can analyse their work).
Target Output: A WIHEA page that disseminates information and possibly diffusion of the actual Warwick AI Essay Analyst tool at Warwick, potentially producing research output
4) Designing Module and Lesson Content & Recommendations for institutional Developments / Purchases: Educator-facing, we review tools and initiatives that might help educators in planning and organising their modules and lessons, as well monitoring their email and forum entries. This group looks at all educator-facing areas besides designing assessments (group 2) and providing feedback on assessments (group 3). This group might also make recommendations to the institution on what software to build or to purchase etc.
Target Output: A WIHEA page that disseminates information, possibly making recommendations for in-house developments / purchase of external software packages
5) A ‘red team’ acknowledges that AI is here to stay and ensures we follow AI Ethics guidelines and that everybody is clear about the risks. This team also reviews and mitigates the challenges to Academic Integrity more broadly. Moreover, it reviews the risk of bought-in products from EdTech and Tech companies, ensuring that AI Ethics is applicable both for in-house and off the shelf, bought-in products.
Target Output: A WIHEA page that provides information for colleagues worried about AIEd (explanations, links to other resources, links to discussions) especially on the topic of AI Ethics and Academic Integrity (what is OK to do, what isn’t – where should students / educators draw the line). Collaborating with stand 1, this group might want to explain (do a high-level literature review / providing links to important research) aspects of AI Ethics / Academic Integrity, such as explaining concepts such as ‘Algorithmic Fairness’. Building on work by other groups, e.g., last year’s ‘Online Assessment Data Ethics Group’, this group might want to develop a proposal for SLEEC (https://warwick.ac.uk/services/gov/committees/sleec/) and/or to provide guidance and advice to EPQ on suitable policy and guidance where appropriate.
Proposed Overall Target for entire Task and Finish Group, i.e. across the five groups / strands: Have some tangible outputs (WIHEA page, blogs, talks) that support colleagues when they embrace change in an ethically sound way that respects all stakeholders, especially learners and educators. Ideally collaborating with other universities, other education providers, and industry. Possibly develop a proposal for SLEEC and/or provide guidance and advice to EPQ on suitable policy and guidance where appropriate.
Please email Isabel.email@example.com if you are interested in joining any of the groups.
Please email Leda.Mirbahai@warwick.ac.uk if you are interested in joining our open WIHEA Diverse Assessment Learning Circle with interesting talks, such as our talks this month on Synoptic Assessments and on Democratising Assessments.
Add a commentYou are not allowed to comment on this entry as it has restricted commenting permissions.