Send message to Authors

Paper short abstract

Multiple (successful and unsuccessful) experiments with AI and OER in India and the UK highlight constraints from students' study practises and access to technology, university policy and resources. We provide reflections and suggestions that might further the OER-AI agenda.

Paper long abstract

The Dubai Declaration emphasises context-sensitive integration of AI in open education. This is an important aim, but faces practical constraints. We reflect on our experiments developing and using open educational resources (OERs) and multiple experiments developing large language model (LLM) based systems using OERs in our economics and marketing courses in India and the UK.

We have been motivated to enable free access for students to quality, critical materials, including for students in rich economies, but very low-income households. Some of those projects (e.g., CORE Econ) have been very successful, others (the marketing project) not so.

We draw on our experiences with multiple projects for reflection and suggestions across three dimensions: 1) students' learning risks (ChatGPT habituation, unsocial learning, uneven access to relevant technology); 2) university resource constraints (content development costs, AI tool evaluation skills and costs); and 3) university policy constraints (IT policies that restrict educators' use of tools, reward systems that de-incentivise development of OERs and open AI/ML tools).

Our experiments forced us to interrogate our educational and technological choices and their underlying assumptions.

They highlight the inadequacy of "one-size-fits-all" AI systems and the necessity of community-engaged development processes. But, they also highlight that without some broader support and coordination, of the type the 2019 and 2024 declarations imagine, such as through the development of open, adaptable platforms, the resource requirements of doing this make it practically impossible.

We take our experience to make a series of suggestions for practically advancing OER-AI integration that furthers the OER agenda.