The term AI is everywhere these days. Suddenly, it feels like every single company frontpage on the Internet says it is either “an AI company” or “using AI to do X”. It is as if the whole world has discovered that computers can do smart things.

But let me tell you what feels like a little secret these days: AI has been around since the 1950s. In 1950, Alan Turing published “Computing Machinery and Intelligence”, introducing the Turing Test and seriously framing the question of whether machines can think. The term “AI” was coined in 1956 by John McCarthy and is widely considered to mark the birth of AI as an academic discipline.

So, what is AI anyway? Artificial intelligence (AI) is a field of computer science focused on creating systems that can perform tasks requiring human-like intelligence, such as learning, reasoning, perception, and decision-making.

Before today’s AI frenzy, I would call AI features “smart” or “data-driven”. It was all AI, just with a different name. Heck, possibly even a better and more inclusive name.

Since at least the 1980s, computers have played Tic-tac-toe, chess, complex strategy games, and more. And during the early 2000s, websites started making design decisions based on what made visitors stay or buy, A/B testing. Much of this was AI and various degrees of “human-like”.

Machine learning is another type of AI. For example, recommender systems (“Customer who bought this, also bought X”), classifiers (spam detection in e-mail), and image analysis (face detection) are all in that bucket.

What is “human-like” is not always that clear-cut. I tend to think of AI not as a binary thing of “Is it intelligent or not? Yes or no.”. Rather, I think of artificial intelligence as a range going from rule-based on one end of the AI spectrum, and on the other end something like AGI – human-level cognitive abilities, enabling it to understand, learn, and apply knowledge across a wide range of tasks, much like a human.

Ever since the 1950s, we as humans have been on a journey along the AI spectrum, in the direction of smarter and more intelligent systems. LLMs are our current last stop.

Whenever someone talks to me about AI these days, I ask what type of AI they are referring to. AI is a long-lived, wide, and deep domain encompassing a lot of things. Not just LLMs.

Finally, just because a company claims they “are AI”, does not mean they are using some mind-blowingly new technology. They might just use tech from the 90s, who knows.