Is generative AI the viral sensation we’ve been promised? Headlines are selling it as a panacea, but 30-year AI industry vet Eric Siegel says that’s mostly hype. It may be impressive and introduce efficiencies, but it won’t run the world as we’ve been promised.

Predictive AI, Siegel argues, holds all of the transformative potential that we’ve been guaranteed with generative AI and the LLMs that have become increasingly popular. While generative AI works on a low-level of detail and isn’t able to be trusted blindly, predictive AI is fully-autonomous, and can be implemented across a wide range of industries.

Here’s why Siegel thinks the actual potential lies with predictive modeling, and why we’re not going to replicate human intelligence in machines anytime soon.

ERIC SIEGEL: There's kind of an illusion with generative AI.

"This promises to be the viral sensation that could completely reset how we do things."

According to all the headlines, it's on the brink of solving all business problems automatically with the slight side effect of displacing huge amounts of the workforce. It seems so amazing. It's potentially a panacea. No. It's hyperbole. It's hype. What we get with generative AI is extremely impressive, but it's not going to run the world. It does have the ability to create efficiencies, but it's more limited. Whereas predictive AI, which is older, very much still has great amounts of untapped value.

I'm Eric Siegel. I'm the co-founder and CEO of Goodr AI, the founder of the Machine Learning Week conference series, and the author of "The AI Playbook: Mastering the Rare Art of Machine Learning Deployment."

I became fascinated with the concept of artificial intelligence as a kid in the late seventies and then in the early eighties. Eventually, my education led me to machine learning, and I've been in the field since nineteen ninety-one.

"Whoa, Kasparov, after the move c4, has resigned."

Now I was sort of semi-horrified with the AI hype for a few decades, and it just got a lot worse in recent years because of generative AI. It's going to feed that frenzy because it's so seemingly human-like. Generative AI, something like chatGPT, a large language model, it is capable of communicating about any topic and often giving responses that seem to understand what you're saying. And I grant that on some level, it has captured understanding and the meaning of words and phrases and sentences and paragraphs. But I believe that the difference between what it can do and what humans can do is going to become increasingly apparent.

Generative AI is sort of correct often only as a side effect. When people say "hallucinate," they're like, "Well, look. It just makes things up." What impresses me is that it actually gets things right sometimes because it's only working on that low level of detail, the per-word level, which results in that sort of seemingly human-like capability. There's a big difference between that impressive capability and the potential value. It's certainly valuable for writing first drafts. So it'll produce a first draft of a letter you need to write or a syllabus or something like that. But you can't trust it blindly. You have to proofread everything that it gives you. That actually, in a way, makes it less potentially autonomous. The whole point of computers is to automate. Right? It does things really fast. And to the degree that we can actually trust it well enough to do things automatically, that ultimately helps the economy. It helps the efficiencies of the world.

Predictive AI, that's the technology you turn to when you want to improve your existing large-scale operation. It does have the potential to enjoy the benefits of autonomy. So predictive AI or enterprise machine learning, that's the technology that learns from data to predict in order to improve any and all of the millions of decisions that make up large-scale enterprise operations. And these are the things that make the world go round. So predict who's going to buy in order to decide who to contact with marketing, which transaction is most likely to be fraudulent to decide which transactions to block or audit, which train wheel is most likely to fail in order to decide which one to inspect. It's not just train wheels. The New York Fire Department does that to predict which buildings are at most risk of fire to triage and prioritize inspections, or which healthcare patient should we take another look at before discharging because they're predicted very likely to be readmitted to the hospital?

All of these predictive applications are a form of prioritization or triage, and the computer is systematically making those decisions over and over again real fast, fully autonomous. So we have data. We give it to machine learning, which is the underlying technology. It generates models that predict, and those predictions improve all the large-scale operations that we conduct. Predictive AI is so applicable across industries. Let's take the delivery industry. UPS is one of the biggest three delivery companies in the United States, and they actually streamlined the efficiency of their deliveries by predicting tomorrow's deliveries. That makes such a big difference that in combination with another system that actually prescribes the driving directions, to this day, UPS enjoys savings of three hundred and fifty million dollars a year and hundreds of thousands of metric tons of emissions.

So this is how it works. When they have to start planning and then loading the trucks in the late afternoon or early evening so that it'll be ready the next morning, they have incomplete information. What they don't know is some of the packages that are still coming in later that night. So what they do is they augment the known information, which is that they already have a bunch of packages in hand that they know are meant to go out tomorrow morning for their final deliveries. And they'll augment that with tentatively presumed predicted deliveries by applying a predictive model for each potential delivery address and saying, "Hey, what are the chances that there'll be a delivery there tomorrow?" Now they have a more complete picture of all the deliveries needed for tomorrow. They can do a better job planning and loading the packages overnight so that when the trucks go out in the morning, they'll have relatively optimal routes that don't take too many miles of driving, too much gasoline, too much time of the drivers. Now some of those predictions will be wrong, but they're confident enough that the completeness now actually overweighs some of that uncertainty. This is what you need to do if you want to improve existing large-scale operations. You need to work with probability. Assign a number. How likely is this outcome?

Here's the thing. It doesn't make a difference how good the number crunching is unless you act on it. It's not intrinsically valuable. The value only comes if you actually deploy it and change your existing operations. We have this incredible seemingly human-like capability of generative AI, which in one sense, I think is the most amazing thing I've ever seen. But underlying the excitement is the idea that we are moving steadily towards and potentially very near AGI, Artificial General Intelligence, which is a computer that can do anything a person can do. It's this feeling of a computer, kind of, coming alive, like Frankenstein, which we see over and over again in science fiction movies.

In the real world, I do not believe we're going to fully replicate humans anytime soon or that we're actively making progress in that direction. That is a recipe for mismanaged expectations, otherwise known as hype. The antidote to hype is simple. Focus on concrete value. Discover whether you're using generative AI or predictive AI. Determine a very specific, concrete, credible use case of exactly how this technology is going to improve some kind of operation in the enterprise and deliver value. If you want to just sort of explore how close is it to the human mind and why you think it might be getting there, that's kind of a philosophical conversation, and that's great. But if you're talking about, sort of, improving efficiencies of operations that make the world go around, I think we should be a lot more practical and less pie in the sky.