It may well be that there are constructive and productive uses for Artificial Intelligence, specifically generative AI such as ChatGPT and the like.
But not in this course. Here, there will be zero tolerance of AI usage.
You will not use AI in this course.
As part of your contract, you sign and certify that you will not be using generative AI, which includes but is not limited to websites and apps such as ChatGPT, Claude, Gemini etc., at any stage or for any reason in your work for this course. You should also turn off any Gen AI features of other programs (such as Grammarly, which you should not be using in any case).
If you use AI, then you are breaking your contract.
You should also read all the texts either in the recommended English translations, or in the original language. You should not read them in any other language, or use translation software. (You may of course use a dictionary to look up specific words or phrases.) Similarly, you should write your blog posts in English, and not use translation software in their production.
If you have any questions about what is permissible or not, you should err on the side of caution and are welcome to talk to me or send me a message with your concerns. I do know that, increasingly, AI is becoming integrated into all kinds of websites and programs, making it ever harder to avoid. But you should do your best to do so, and we can discuss limit cases or grey zones.
A contract depends upon trust and transparency. If you are having any difficulties at all in the course, for whatever reason, please do feel free to talk to me about them.
Why am I so strictly opposed to AI in this course?
Traditionally, assessment focuses on product: on what you are able to achieve after you have done something else. For instance, you write an essay or take an exam to show what you have learned from studying course materials, and how well you do in that essay or exam is taken to be a reflection of what you have learned, and what you are now able to do (progress, for instance, to a higher level).
By contrast, assessment in this course focuses on process, rather than product. What counts here is reading, in all its difficulties and complexities (if also pleasures), rather than whatever conclusions or insights you may or may not draw from that reading. Similarly, what counts is that you have engaged and have a perspective on the texts, rather than whether or not your perspective is the “right” one. What counts is taking part, rather than achieving any particular goal.
AI is pretty good (if still not perfect) at coming up with a plausible product. But its use eliminates, to a greater or lesser extent, the process that is important here.
What is my view on AI more generally?
AI is evolving and changing very rapidly, as is its place in our society and everyday life. It can be very seductive: it seems to make things so easy. But it is worth stepping back for a moment and considering its strengths and weaknesses, without ever forgetting also its extraordinary environmental impact.
- It is unreliable, most obvious in its so-called hallucinations (aka “making shit up”). It is getting better at avoiding these, but the people who engineer AI programs are still not sure why and when these hallucinations occur, so they are unlikely to disappear entirely. AI’s guiding principle is plausibility, rather than truth.
- It forestalls novelty and creativity. Programs that draw, for instance, on Large Language Models work by combing huge databases of what has previously been written or said, and only ever repeat what they find, as they construct new statements that are likely or plausible. AI works through statistical probability of what is likely based upon the past, rather than seeking to come up with anything that would be truly novel.
- It is relatively good at bureaucratic tasks, or more generally tasks that we simply do not see as having any inherent value. Indeed, as soon as we use AI to complete a task, we indicate that we do not think that task in itself is worth doing.
- It is also, frankly, fun to play with.
So long as we do not take AI seriously, and treat it solely as either a game or a drudge, then I feel we are on the right track.
I dream of a day in which we have a reliable and effective AI to which we can consign the tedious, semi-intellectual bureaucratic labor in which we see no inherent value: filing taxes, writing reports, filling out forms, university admin, and so on. This would free us up to do what is interesting and valuable in its own right, which for me includes reading, thinking, collaborating with others, and having conversations about what we have thought and read (i.e. what we do in this course), and for others may include, for instance, making art, writing stories, playing sport, and so on.
Sadly, at present it looks as though it is more likely that we hand over the interesting stuff to AI, leaving us still with the bureaucratic labour of endless bullshit jobs. Believe me, this is not a future that you want.
It is not too late, however, so long as before we outsource a task to AI, we ask ourselves what would be lost by our not doing it.
For more thoughts, I find British computer scientist and critic Dan McQuillan, author of the book Resisting AI, is provocative and informative.
