Resolving a Wickedly Sticky Problem

Sticky problems

One thing I was seeking to find out at IATEFL this year was how others are approaching the sticky issues around artificial intelligence (AI) in university language learning settings. In my current job, the problem of how to distinguish genuine student work from an AI produced piece of writing is proving to be a "wicked problem".

A wicked problem was originally defined as "a class of social system problems which are ill-formulated, where the information is confusing, where there are many clients and decision makers with conflicting values, and where the ramifications in the whole system are thoroughly confusing’ (Churchman, 1967, p. B-141 cited in Termeer et Al., 2019).

I feel this describes the situation with AI mis-use in the higher education setting at the moment. It is a systemic problem born from an ill-conceived idea that everyone can access a powerful, artificially intelligent tool despite not really understanding how to use it effectively without prejudice, and which institutional leaders have little sense of responsibility for controlling its usage other than to ban it completely.

The fact that among the wealth of experienced language teaching professionals none of us can offer a definitive solution to this problem suggests this is a sticky problem that requires further consideration and exploration.

I have heard from colleagues who are applying innovative and creative measures to getting students to engage honestly and transparently with AI tools for writing tasks - in particular business English centers teaching students how to produce authentic texts that meet requirements of register and contextual appropriacy.

Collectively, the message I am getting is that we need to focus on how to assess the process rather than the product. We need to be taking a hybrid approach and leveraging the AI tools to act as support for language learning and that students need to be curators of the process - constantly questioning, checking and revising not only the AI output but also their own prompt engineering skills.

With another day of conference talks to listen to, I shall continue to listen, learn and contemplate how to foster the best learning for my own students with help from AI rather than find ways to prevent learners from delegating their learning tasks to a bot.

Reference:

Termeer, C. J. A. M., Dewulf, A., & Biesbroek, R. (2019). A critical assessment of the wicked problem concept: relevance and usefulness for policy science and practice. Policy and Society, 38(2), 167–179. https://doi.org/10.1080/14494035.2019.1617971