Shall We Play a Game?

 Thanks to the hard work of the Tech Rav I have been informed that there is an active chatbot which uses artificial intelligence and can do things like parse code, hold conversations and write stories and essays.

I'm not jazzed. Let's discuss.

I'm in favor of intelligence. I think one of the smartest things we can do is be intelligent. And I'm even OK with aritifical intelligence. I think of it like artificial ingredients -- as long as the food tastes good and can last in the car's glove compartment for an extra coupla weeks, then I'm good with it. The problem is in the impact (and I'm turning teacher now) on the educational process and the development of the adolescent brain. When applicable.

So, first thing -- I tried it out, and saw the results of others' attempts to use the service. My assessment is "good but not great." The chat service has limitations built in and it knows it so it admits those at every opportunity. It is very careful about explaining that it cannot think or opine (though it insists that it can synthesize and innovate, claims I find hard to test as it would require that I know all extant conclusions and can spot if the AI provides something heretofore unrecorded), and somehow is programmed to avoid (whatever the programmer's agenda defines as) hate speech or speaking too admiringly of people that it is not in vogue to admire. It ends up in a circular loop when you ask it about how it assesses sources it uses. But it is a very powerful natural language research tool. The essay/book summary I read was pretty poorly "written." I have to wonder what sources the AI borrowed from in order to produce the content (as I doubt the computer actually read the book in question), but the sentence structure and word choices were not good at all. I say this with a caveat -- to a casual reader, I would guess that the sentences sounded fine. But to an English teacher, or one really inspecting the structure, they were below average in composition.

Years ago, I envisioned a scenario in which a computer wrote poetry. The masses loved it (the pretend poem, not the idea) and this was the saddest part. I also saw a TV program on which an elephant painted a picture and a child created something which passed (to the uninformed) as a Jackson Pollock. The line between brilliance and crap seems awful thin sometimes. Art is always subjective so if an AI wants to create, I can see that creation passing all sorts of tests and vetting because the end result either resonates or doesn't, source be damned. The fact that a text is not driven by actual experience and insight might be troubling but if we calue the end product over the process, then so be it. I did try out one of those AI-art websites at which you describe the image you want and the system creates it for you. It didn't. Major fail so I'm still in need of an illustrator for a side project. That's a story for another day.

But over the years, I have heard, too often, about skills which are no longer relevant because we have technology which can replace them. How many people think that any of the following is still worth teaching (or learnng)? Note -- this list is not about physical skills but intellectual exercises:

reading maps, doing basic math, using a library/research, interpolating logarithms, reading an analog clock, learning to spell

Odds are, you saw at least one item on that list that you don't think we should spend time teaching because of the ubiquitous technology which makes that no longer a valuable skill. So, next question -- would you be ok if any of the items on the following list fell by the wayside in the same way?

driving a car, writing an essay, reading a lengthy book, diagnosing an illness, doing advanced math, designing a bridge, umpiring a baseball game, providing emotional therapy, ruling in a court case

Some of you would be more than happy to hand any and all of those items over to the ersatz mind of silicone. Others might try to draw a line somewhere and explain why "yes to this but no to that." Hours could be spent arguing but that's not my point. That's just a hobby...

Here is where I see the problem -- NOT in that a student might employ an AI system to write an essay and then said student might hand in the essay pretending to be the author. Yes, that's bad and, yes, teachers will have to find ways to detect artificially constructed material. But the big problem will be when people say that teachers DON'T have to detect constructed material. At a certain point, I suspect that people who say, "It is fine that no one has to learn how to read a map because we have computers  which hcan tell us turn-by-turn directions" will say, "It is OK that no one has to learn how to write an essay because we have computers who can do that just as well." Those people insist that we are freeing students up to apply themselves to more complex thinking and learning because we aren't bogging them down and wasting their time with easy stuff that technology can do for us. And to the concern that the AI created responses are not very good (and if a student were to turn one in, it would be cause for learning and inmprovement, not as a static end result) too often people rely on the "but it is good enough so since it seems like a student could have written it, it shows that a student doesn't have to waste his time writing it."

What people don't understand is that knowledge is a progression and the advanced and critical thinking skills that we all say we value only exist on a solid foundation of basic skills. All those higher order thinking skills cannot be built on staircases of sand. One cannot play chess on an advanced level without learning the basics of the game and memorizing how the pieces move. One cannot be a successful NBA player without learning the basics of how to dribble. All advanced knowledge and all advanced skills come as the result of long term hard work, standing on the shoulders of practice at the fundamentals. If we cut the legs out by saying a student doesn't need to know how to do addition in his or her head because he or shee will always have a calculator nearby, then not only are we disadvantaging that student when he or she is in a tech-free environment (yes, they exist), but we are not demanding that the student understands the mechanics of arithmetic which will allow for the model of thinking which unlocks advanced math. Not knowing "how" leads to never understanding "why" and that presents a problem. A student who lets a computer do research never learns the value of the hunt, the struggle of thinking unpredictably in order to find something unexpected, or the need to evaluate sources and prioritize data. A student who lets the computer write never learns to organize thoughts, or communicate persuasively, or work reflectively and struggle to improve -- there is no respect for the rules of grammar so all communication becomes tainted by ignorance. These are exportable skills which can't really be taught in a vaccuum. They are part of a brain-growing mindset. How can we expect that we will have a next generation of advanced thinkers if we don't demand the development of thinking skills but instead, assume that at a certain age, never having had to grow, those brains will be ready to be advanced?

I'm no luddite. I'm not even ludesque but I like my pedagogical change like a like my roller coasters -- slow and predictable, without sudden changes and with not a lot of vomitting. AI has a place, and, yes, that place will grow as we figure out the most effective way to incorporate the advantages of machine thinking without losing the elements necessary to keep human thinking staying steps ahead, so that we can design the next generation of AI.

Meanwhile, I, a human wrote this. So there.