This is the first part of a two-part blog series on the disruption of AI. Catch up on Part II here.

I recently read a New York Times article about the diffusion of AI and wanted to share a few words about it from a disruption angle. The article begins by saying that AI doesn’t have to be perfect to be useful and uses some examples like people using phones to take pictures instead of high-end digital cameras. This comparison insinuates that AI is a new innovation. It’s not.

A Quick Review of Diffusion

We have discussed in previous blogs how innovations diffuse through stages with different types of users/consumers. It begins with the innovators, remember? The people who are a little crazy and very few people listen to them. They are the first ones to see the value of a new idea and put it to work. They are friends with conservative, respected, early adopters who sift through the innovators’ many ideas/gadgets and adopt only those they find useful and reasonable for their communities. Early adopters begin the upward sweep of the innovation’s bell curve of diffusion by engaging the early majority. The timeframe of diffusion could resemble an S curve as in the rollout of the car or an R curve as in the rollout of the radio or iPhone. As more people adopt the innovation, money flows in from sales, and the innovation, which might have started out clunky like the first digital camera, radically improves and eventually overtakes the capabilities of the disrupted device rendering it obsolete.

AI Diffusion

AI has, of course, progressed through many of these stages. Having started as a Neural Network with Dr. Gregory Hinton back in 1972, and although this other recent New York Times article seems to represent it as rather early in its life (still clunky), it is in fact far along. I have been in technology for decades and I remember the first time I heard about Artificial Intelligence. It was in the late 80s when I was on the General Motors team with IBM in Detroit. It is probably safe to say that IBM was an innovator with the technology and introducing it to some select customers. These customers would have been the early adopters. We are WAY beyond early innovations with AI.

Think about how we as everyday consumers use AI. “Alexa, turn on the living room lights.” “Google, what is the weather today?” AI is fine-tuned and assists many of us every day through the talents of ingenious programmers that work for some of the biggest technology companies, like Apple (Siri), Amazon (Alexa), Google (Google AI) and IBM (Watson). The difference between the AI that we use every day and the newest release of ChatGPT is that we have moved out of the proprietary world of development into the mass diffusion of cloud based open source. This latest diffusion should offer AI to the late majority and beyond, especially those businesses that cannot afford the programmers that make more than NBA players.

ChatGPT, while it might appear a “little clunky” (if you’re an AI programmer), is an elegant option for businesses vulnerable to AI disruption. Remember, disruptive innovations do not serve the entrenched users (Apple, Amazon, Google, IBM), they serve a new community of people who are looking for a different, less complicated, cheaper way of doing something.

Let’s use the New York Times example of the iPhone camera users vs the high end digital camera photographers, which is actually a repeat of when digital cameras took over the market from 35mm. The iPhone cameras were not designed for photographers. They were designed so that millions of people could take pictures in a convenient, inexpensive way, because they are not buying an expensive camera and hauling it around. As Apple realized how much people liked their always available cameras and not so famous people were taking amazing photographs, they improved the camera to the point where ordinary people have an extraordinarily sophisticated camera in their pockets. Now, can I take pictures as well as a professional photographer? Of course not. She understands light and backgrounds and colors and I am just snapping pictures of my dogs. But, are my pictures a thousand times better than they ever were? Yes! My dogs have never looked so clear, and every once in a while the iPhone will have the perfect conditions. I won’t have smudges all over the lens and I will get a perfect picture of a sunset, or my grandson or husband. That is a huge value to me. Clearly, we don’t have to spend any time discussing the success of the iPhone camera.

AI’s success is ubiquitous as well. As I write this blog, Microsoft is making suggestions for better word choices and the correct use of commas. My iRobot vacuum is figuring out how to get back to its base from the master bedroom. That is all AI. Using cloud platforms like ChatGPT, an ordinary business can now incorporate AI into their applications. Will they be able to program as well as the Google or Amazon guys? Not at first. The big guys have been programming with AI for years and developed their own proprietary “platforms”. But now ordinary businesses can incorporate AI into their applications to transform ideas into products inexpensively. The New York Times says that ChatGPT is imperfect, makes mistakes and struggles with more complicated tasks or programs. Maybe, but again, the audience is not Google and Amazon, it is ordinary businesses who are not doing anything sophisticated right now. And as money rolls in from early ChatGPT users, it will get better and eventually, it will exceed the capabilities of many existing programs. 

AI Consequences

The significant part of the story is the section on unintended consequences. This doesn’t even scratch the surface. When we look at consequences of disruption we must look well beyond the surface as the New York Times article indicates. In their example of a political cartoonist who asks ChatGPT to describe a relationship he has with another cartoonist, the bot makes up a story that isn’t true. That is incredibly superficial. Like when the birth control pill was introduced and women had fewer children. The real changes were vast. Women were able to work longer before having children, if they had children at all. They were able to make more money and achieve more success and that started to impact the traditional gender roles, and so much more. That was the real impact. Think Butterfly effect or Chaos theory.

What will be the impact of ChatGPT? Stay tuned!