AI-generated essays - time to rethink written assignments

Students will employ AI to write assignments. Teachers will use AI to assess them. Nobody learns, nobody gains. If ever there were a time to rethink assessment, it’s now.

This is a quote from an article by Mike Sharples, London School of Economics, New AI tools that can write student essays require educators to rethink teaching and assessment. There are now tools that use artificial intelligence (AI) to generate highly plausible academic essays, complete with references. The article takes an example of a short essay about the problems around the popular concept of learning styles generated by a Transformer AI program, GPT-3. The user simply entered the first sentence and AI completed the essay. The result is not particularly insightful but good enough to pass and it won't show up in any plagiarism control since the text is completely original. If the essay is fed back into the tool by the teacher it can write a similarly plausible comment on the essay. The whole assignment can therefore be performed by AI, prompting the quote above.

Even if some of the essays generated this way may still have weaknesses (the example in the article has false citations) the whole point of AI is that it is constantly learning and improving. The phenomenon is not new, it has been possible for many years to pay someone else to write your essays for you via essay mills, but now the human element has finally been removed. Does this mean the end of the written assignment as an examination form? 

The author suggests a few ways of using the AI transformer in a constructive way, for example by getting students to generate AI texts and then find faults in them and improve them. This reminds me of how some teachers tackled plagiarism by writing a sample essay/article that included several form of plagiarism as well as poor citation practice and asked the students to find the problems and correct them. 

But the main point here is that we need to move on to new ways of assessing students and avoid examination methods that ask questions that can be automatically generated or copied from the internet. 

Finally, as educators, if we are setting students assignments that can be answered by AI Transformers, are we really helping students learn? There are many better ways to assess for learning: constructive feedback, peer assessment, teachback. If Transformer AI systems have a lasting influence on education, maybe that will come from educators and policy makers having to rethink how to assess students, away from setting assignments that machines can answer, towards assessment for learning.

Live assessment activities like interviews, presentations, debates or round table discussions can be run either on-site or online and are almost impossible to cheat in. But that brings us to the eternal question of how to move the focus in education from extrinsic motivation (exams, credentials) to intrinsic motivation (satisfaction, self-confidence, pride). Focus on competition, rewards and results encourages cheating among some, whereas activities that focus on community, learning for pleasure and intangible rewards are generally free of cheating. If learning is in the forefront there is simply no point in cheating. Hopefully AI-generated essays will remain a mere curiosity.