There have been numerous news reports recently about
ChatGPT. This is the new chat interface based on a type of neural network that
gives human like responses to natural language input. I am sure you all knew I
would need to address this at some stage, but I was reluctant as I suspected it
would present Nick with too great a temptation to question my humanity. You can
try ChatGPT yourself at https://openai.com/blog/chatgpt/.
For those who would rather not, I have posted a transcript of a discussion I
had on the relationship between “theological reflection” and “reflective
practice” to https://example.commonawards.org/mod/page/view.php?id=1571.
This includes factual errors (mistaking Stephen Pattison for Elaine Graham)
when I asked it to generate a bibliography, although it would give me the
bibliography using any referencing scheme I wanted. It is however very
impressive. The challenge for education from this type of tool will be huge.
However I think it will not be the obvious issue of students getting ChatGPT to
write their essays. This can be fixed by more diverse forms of assessment
(conversations, group projects and perhaps exams). Furthermore, plagiarism
detection tools will themselves become smarter – apparently AI systems are very
good at detecting whether a piece of writing was generated by an AI system.
What is more challenging is how you teach people to work with such tools. Tools
that can take a page of your hastily written thoughts and convert them into
polished prose in a particular style. Or even take a sermon text and convert it
to a different style of preaching. Yes I experimented with this, yes it worked
frighteningly well and no I am not going to share the resulting sermon text,
nor will I preach it. Finally, for avoidance of doubt, I don’t use ChatGPT to
help write these pieces, but the question is, if I did would you see an
improvement and would you care?