- This HEPI blog post was kindly authored by Ninian Wilson, Press Officer at the Quality Assurance Agency.
It’s been a busy time for QAA responding to the rise of generative AI. After three webinars attended by over 2,000 people, discussions with providers across the UK, a brief trip to Tbilisi to share best practice with European quality agency colleagues, one piece of guidance out and another to be published imminently, we have been reflecting on the perspectives of the technical experts, academic staff and students that we brought together. So, what have we learned?
We can’t – and shouldn’t – ban it
All our event speakers agreed that banning generative AI in tertiary education is not only unfeasible but undesirable, including in assessment. Detection tools lag behind the software itself – we will shortly see this technology embedded in tools we use every day and being able to use it effectively will become a desired skill in the labour market.
That said, we need to avoid scenarios where students and learners are using AI to churn out work without effort or critical thought. In response, we should help students understand the limitations of these tools, appreciate the negative effect that overreliance can have on their own learning experience, and focus assessment on the learning process, not just the output.
Don’t panic
This technology, the change it poses to higher education and its implications for academic integrity are not wholly new. During the first of our webinars, Dr Bronwyn Eager likened the emergence of ChatGPT to a fish asking ‘what is water?’. From spell check to learning analytic platforms, AI has been the ‘water’ all around us for some time. It’s the open access element that is new.
Much like the advent of the printing press to typewriters and calculators to Google, what start as perceived threats to the status quo quickly become the new status quo. Generative AI tools are the latest of these milestones and Bronwyn encouraged us to see them as a powerful tool rather than an existential threat.
There have always been grey areas in academic misconduct – a parent editing an essay, too much peer exchange, taking inspiration from existing essays online, etc. Generative AI exacerbates this, so it is more important than ever to promote academic integrity amongst students, engage them in authentic assessment design and view uncritical AI use as a symptom rather than cause of a much larger issue.
Access and training for staff and students
With many generative AI tools sitting behind paywalls, our event speakers felt it will become important that providers enable equal access to these tools to prevent more advantaged students benefitting from the technology compared with less advantaged students.
And given how ubiquitous generative AI is going to become, there was discussion of how providers will need to rapidly bring their staff up to speed in AI literacy and skills to support students to use it critically and ethically. Much like the existing academic induction students receive at the start of their courses, using generative AI tools must feature as a prominent induction or indeed standalone module.
Re-thinking the paradigm
While providers respond to the immediate environment, our webinars pointed to a whole stream of higher-level strategic thinking the sector needs to do. One of the most compelling questions that Sarah Eaton posed in her 2021 book, is whether we are entering a post-plagiarism era. With human / AI writing shortly to become the norm, what do we mean by plagiarism? How long will it be before institutions asking students to declare use of generative AI tools in their assessments looks as futile as asking students to declare use of Google? How do we ensure students credit ideas that aren’t their own if the idea was generated by a random statistical distribution of words?
For the full picture on these discussions on ChatGPT, check out what the experts had to say in our ChatGPT webinar series. QAA’s guidance for institutions on responding to generative AI is also publicly available on the ChatGPT and Artificial Intelligence page on the QAA website.