Today I am not at the International Medieval Congress, but as mentioned on what was supposed to be a holiday but is actually looking after ill cats. It’s OK. It also gives me time to put together a lighter-weight distraction blogpost in order to make up somewhat for the lack of blogging you have recently endured from me. So here we are. You may remember that part of the reason for that lack was that I got myself into a sunk costs problem over writing a long-distance review of a book called Hacking the Academy, which was a lot of work. Given how much work it was, it was tempting to find a way of outsourcing that work, and since we are in the middle of what might most gently be called a phase of adjustment to the academic implications of ChatGPT and large language model software, I couldn’t help wondering what the so-called AI would make of the task…
Advertisement banner image for GPT-4 from Matt G. Southern, OpenAI Releases GPT-4: Now Available In ChatGPT & Bing
, Search Engine Journal, 14th March 2023, linked through
This was really my first brush with the beast in its lair, and I drew the line at signing up for an account. The career blackmail possibilities of their user data must be tremendous, and I saw no need to add to them, so all they have on me is an IP address. Well, and this post. But without an account all you can use is the online chatbot here, so I did. And these are the results.
Screenshot of interaction with the OpenAI ChatGPT chatbot, reading as follows:
AI: Hi? How can I help you?
You: What do you think of the 2013 book Hacking the Academy?
AI: As an AI assistant, I don’t have personal opinions, but I can tell you that "Hacking the Academy" is a collaborative volume that explores the potential of digital media to transform academic institutions. The book includes a collection of essays by scholars and educators who offer innovative ideas and strategies for using technology to improve research, teaching, and institutional practices. If you are interested in this topic, I can provide you with more details or recommend related materials.
You: Do people think that the 2013 book Hacking the Academy still has value?
AI: Yes, many people still find value in “Hacking the Academy” even though it was published in 2013. The book’s focus on the potential of digital media and technology to transform academic institutions is still relevant and important. Additionally, some of the essays in the book offer insights that have only become more relevant over time, such as the challenges of academic labor and the need for greater diversity and equality in higher education. Overall, “Hacking the Academy” remains a thought-provoking and valuable resource for anyone interested in the intersection of technology and academia.
You: Thankyou!
Error while calling OpenAI: rate limit reached for default-gpt-3.5-turbo in organization org2MYfTx8Pgtl3C8H3nS3EjVU2 on tokens per min. Limit: 90000 / min. Current: 89789 / min. Contact us through our help center at help.openai.com if you continue to have issues.
I have far fewer issues than I expected, in fact, though it is ironic that what looks to be an unconnected system error arose at the point of purely social exchange. The repetition is noticeable between the two answers but so might it have been in a human’s answer. This also isn’t really a full test of the thing’s possibilities, I do realise; the second answer is really the outline of a fuller review which would, in the hands of clever users, have expanded on the second and third sentences until you had full paragraphs for each point written by the robot. 1,000 words would have been pretty easy to get like this. But there’s also nothing really wrong with the substantive bits of what it told me. It found information on the Internet and made as reasonable a summary of it as anyone could. Obviously I could make it work harder, but the fact that it’s capable of doing what we want but not acting as we do is still fascinating. (We should also note that that error message suggests that chatbot is still running on GTP 3.5, rather than the more capable GPT 4.0, so if I’d dared the account I might have got my post much more easily.) I’m sure I’ll be dealing with this stuff more in future, but still: in a very simple and direct way it’s going to be difficult to make friends with it…
A Corner of Tenth-Century Europe 