There is this joke that people do statistics in R, machine learning in Python, and AI in PowerPoint. (I did not say it is a funny joke.) The point is that AI is not only about technology and programming, but also about framing machine learning in business-friendly terms to convince tech and non-tech savvy audiences alike about the use of cutting-edge solutions. That is, what we call AI can sometimes be seen as the combination of machine learning and marketing.
There is a lot of hype surrounding AI, and rightfully so. ChatGPT and other services help me on a daily basis in ways that cannot be reduced to “advanced word guessing”. This success leads me to the counterintuitive conclusion that AI could be dead in five years. I do not mean that we currently are in an AI bubble (although that could be the case), or that the current use of AI will slow down (which I find unlikely). On the contrary, I doubt we will talk a lot about AI in five years because AI will be an integral part of how people search for and interact with information. Or, if we talk about AI in five years, it will most likely refer to different technologies and solutions than what we denote as AI in 2024.
I doubt we will see a lot of products talking about being powered by AI. More specifically, here is my working hypothesis: machine learning equals AI plus time. Years ago, before we talked about large-language models and multimodal learning, we talked about features like the Netflix recommendation system and chess-playing engines in the context of AI. Take, for example, this description in a blog post from 2020 about Netflix and their recommendation system: “When watching for some time, Netflix can predict and recommend what kind of movies we are likely to watch in the future. The more we watch, the better the recommendation. This is one of the applications of Artificial Intelligence (AI), a recommendation system.”
The point is not that the above quote was incorrect at the time, or even incorrect today, only that people usually have other things in mind when they talk about AI today than what is being recommended on Netflix. Similarly, I already see people talk less about ChatGPT as AI than a few years ago. The more popular AI will be, and the greater the hype now, the less likely it is that we will talk about AI-powered services in five years.
This reminds me of another blog post from 2020, written by Lewis Gavin, on how big data will be dead in five years (hence, the title of this blog post). I remember all the hype about big data. It is not that we no longer talk about big data because it is irrelevant. On the contrary, we do not discuss it because it has become ubiquitous (and most AI now is fueled by big data).
This is not to say that we will not talk about AI at all in five years, in the same way that we still talk about big data today, only that we will not talk about AI in five years because of its success.