Artificial intelligence (AI) is the topic of the day in some quarters. Computers guiding self-driving cars is probably the most well known example. They have to “think” in a way as the best human driver would. Too far there are enough glitches to keep them from being trustworthy, but that’s likely to change soon.
Siri, Alexa, and Google Maps give the impression of having artificial intelligence, but they are really just sophisticated programs that can retrieve a limited amount of data from a data bank, delivering audible answers to simple spoken queries. More impressive is the new Chat GPT app that can write papers on almost any subject, imitating the style of a human author. My friend Richard and I messed around with it to see what it could do. Richard is a law professor so he asked it to write a court opinion on a particular classroom case, and I asked it to write a short sermon on The Good Samaritan.
The court opinion was “not bad” according to Richard, but about half the citations were inaccurate. Ding! Mistrial! The short sermon proved better. It would have satisfied most congregations. Everyone would’ve gone home feeling good about it. On the other hand, I was deeply dissatisfied. It was a stitched together mosaic, very sewn, of every Good Samaritan platitude ever inserted into a sermon. A homiletics professor would have torn it apart.
With other teachers, I am skeptical. It’s free, at least for now, and easy to use. I could ask it to write an essay on the philosophy of David Hume, and it would turn out a decent product, bland but decent. That’s great, but I would not know anymore about Hume or his philosophy than I would had I read an encyclopedia article, and probably not that much. I would not have read Hume, reflected deeply on his words, and written an insightful reflection essay. I’m an old guy. It’s been many decades since I’ve read deeply into subjects in my fields of study. So yes, I find myself refreshing memory with Wikipedia articles, but at least I’m studying them and reviewing citations.
Richard was less skeptical. With a little tweaking he thinks Chat GPT could become a useful tool for producing rough draft opinions, particularly in appellate courts. It would relive clerks from hours of sifting through relevant case law to find the right avenue for leading to the right final opinion. Maybe, but I think more than a little tweaking would be needed.
We both speculated about whether Chat type AI would be useful to physicians who needed to know if a set of symptoms might be related to some obscure disease far outside the norm or easily available sources. In any case, AI is here, it’s not going away and it will become more sophisticated and more pervasive.
Chat usage means serious questions need to be asked now about ethical boundaries, consequences for human learning and wisdom, impact on education, and the consequences of surrendering human effort and judgment for a computer’s supposed greater expertise. Great leaps in knowledge have come from informed intuition by persons who have thought deeply about the problems of knowing and understanding what is not yet known or understood. Philosophers in every century have probed the meaning of things. To many it’s looked like a useless waste of too many words, but every generation has benefitted from through better, more just governments, improved relationships between people and nations, and a more clear understanding of what it means to be travelers on this fragile earth, our island home. Can all of that be surrendered to AI? Should it be? What would become of humanity? Does it matter?
A very old episode of the original Star Trek explores a world that had surrendered to AI. It made life easy, pleasurable and prosperous for all. Was it a good world, a desirable world? It seemed like paradise itself, at least for a short time. Then it turned to “hell” for beings that treasured adventure, challenges, excitement, and the pleasure of succeeding at a task that required their full intelectual effort.