Well I am a late comer to the recent onslaught of AI and over the last month have used Bing’ s AI as well as something downloaded from the Apple Apps store called AI Chat. I have also been using ChatGPTZero to scan essays and assignments to determine if the text is "likely to have been generated by humans." Based on my experience, I would not say we should be dispensing with humans just yet.
Students are using AIs like ChatGPT to write their essays and
assignments but it is a mixed blessing for them. As one of my submitted economic history assignments intoned, “NAFTA
was signed in 1945 by Canada, Mexico and the United States.” While AI can be a useful tool for outlining
essays and projects, ultimately the student requires a core of knowledge to
evaluate what the AI is providing. If we ask AI questions about
subjects we know nothing about, we run the risk of repeating substantial
factual errors.But then, that also applies to interpreting the results of any search engine.
The best way to evaluate the ability of an AI is to ask it questions about something most of us know a fair amount about. Case in point, on a whim I asked one of my AIs to compose a retirement announcement for myself summarizing career highlights and achievements as well as some publications. This is the question I phrased: “Write me an announcement on the retirement of Professor Livio Di Matteo, economics professor that summarizes major career achievements, honours and publications.” Here is what it came up with:
Dear Faculty and Staff
It is with mixed emotions that we announce the retirement of Professor Livio Di Matteo from the Department of Economics at our esteemed institution. Professor Di Matteo has made an incredible impact on the economics department, and his departure will be felt greatly.
Throughout his impressive career, Professor Di Matteo has achieved numerous accolades for his contributions to the field. He was appointed the first Jarislowsky Research Chair in Economic and Cultural Transformation in 2013, and has also been the editor of the Canadian Journal of Economics from 2012-2017, among many other notable achievements.
Professor Di Matteo has published extensively, with over 50 scholarly articles and book chapters to his name, and with his research focusing on areas such as health economics, public finance and transportation economics, he has made significant contributions to the field.
In recognition of his contributions, Professor Di Matteo has received many honours and awards. He was elected as a fellow of the Royal Society of Canada in 2015, and also received the Distinguished Brock Scholar Award in 2016.
As we bid farewell to Professor Di Matteo, we would like to express our gratitude for his tireless dedication and contributions to the field of economics. His legacy will remain and continue to inspire students and faculty. We wish him all the best as he embarks on this new chapter of his life.
Sincerely,
The Economics Department
Wow. All I can say is this is not the career I remember aside from some publications in health and public finance. No mention of anything I have done in economic history and I am sure it will be news to the editorial board of the Canadian Journal of Economics that I served as journal editor from 2012 to 2017. And given I was apparently made a Fellow of the Royal Society in 2015, I have yet to receive a letter informing me so. And what is a Brock Scholar Award? More odd, no mention of the university I have taught at over the last 30 years.
What to make of all of this? Has the AI mixed me up with someone else? But how many economists named Livio Di Matteo can there be? I think users of AI as a sort of Google on steroids need to be pretty cautious - caveat emptor is the appropriate expression. Is an AI Chatbot useful? Sure in a general and preliminary sort of way but you need to double check everything it tells you as fact.
In the case of my AI generated career highlights‚ perhaps there is either something amiss in how the algorithms interpret questions or perhaps AI transcends the multiverse of time and space and borrows bits and pieces of alternate versions of reality and spins them into a unique narrative of what might have been? Unlikely of course except as perhaps a plot for a good Sci-Fi story. Yet the question remains, where did all this stuff come from? Some type of random compilation of academic achievements that sound good? Obviously, AI is just in its infancy. Nevertheless, it can be useful as a search and research organizer. It can even be entertaining as in the cases where I got it to write poems about economists. And it was useful in suggesting code for commands or STATA estimation routines. No doubt the best is yet to come. But for the time being, a lot of caution is warranted.
