AI Use Case: Dr. Tony Frontier on Caitlin Tucker's The Balance

This post draws heavily on Dr. Tony Frontier's appearance on Dr. Caitlin Tucker's podcast, The Balance, the full episode can be listened to here.

Having presented multiple times to diverse groups of adult learners on AI - the basic structure of my presentation remains the same - roughly answering three questions:

1. What is AI?

2. How do I use AI?

3. Should I use AI?

This first question seems evident to me as a necessity for those of us in the education field. Dr. Frontier rightly pointed out an essential truth about A.I.. Dr. Frontier draws from the magician, Teller of Penn and Teller, relaying an idea that is also expressed in this interview [emphasis added]:

Matt: “So why don’t you explain all your tricks?”

Teller: “Because the short explanation—the explanation that you’d have to do during a theatrical or TV performance—is dull and no fun. The greatest secret to making a deceptive piece of magic is you do it by the ugliest possible means. It’s complex, it’s unromantic, it’s unclever. Because there are no big secrets. There is no safe full of magic secrets somewhere. Jim Steinmeyer said he thinks most of the public believes there’s a big safe that contains all the magic secrets. The biggest job for a magician, he says, is to conceal the fact that that safe is empty. Because every magic secret is just a minor modification of something that you fully understand in everyday life. Take suspending something with a thread, for example. Everybody’s not been able to see a piece a thread when they were trying to put it through a needle. What makes it difficult to find is lighting and background. If a magician’s using a thread on stage, say, to levitate a ball, he must use lighting and background to conceal the thread. There’s no obscure secret in that. You learned that playing in your grandmother’s sewing box. Every magic ‘secret’ is hiding in plain sight in the everyday world. It’s not news, and eminently drab.”

Matt: “So you think people would be disappointed by the explanations?”

Teller: “Disappointed and bored. On the other hand it’s great fun for an audience to fantasize about the romance of magic secrets..."

A.I., Dr. Frontier points out, seems like a magic. We enter something and it spits out text, music, art, video, or some combination in seconds. The generation is opaque and happens behind a white chat interface (for most bots). For the uninitiated this seems like a Delphic oracle - and something that could become more and more isolating (a note for a future post) as character AIs replace real friendships and relationships.

However, it is not magic. It is simply a prediction of what best answers a prompt originating from millions of training points and refinement of a model to produce a desired response (which may or may not be accurate).

We can bring the magic down to earth - yes - even make it "eminently drab" by explaining the process behind it and thus elucidating the potential follies and promises generative A.I. has for us. In fact, Frontier points out that this is the essence of education - to take the unknown or magical and show the learning process - to build up a sense of what a thing is and how it works so that it is understandable and truly usable. This can apply to a skill, concept, or fact.

"Hey ChatGPT, can we get a graphic showing how AI produces output based on training data?"

Frontier also helpfully offers four starting maxims, which I'll share as well:

1. Lead with learning - adults need to explore before they guide others

2. Take your time - be mindful, ask questions, and don't feel like you need to rush

3. Integrity - be human-centered

4. Fidelity before efficiency - have clarity of purpose; A.I. shouldn't just be an add-on, the education space is overwhelming as is and efficiency is not efficacy.