Archived: Stop Calling Everything A.I. It's Just Computers Doing Computer Stuff

There’s a really interesting lesson here for every business because, on the one hand, the technology we’re talking about is pretty incredible. On the other hand, however, it really seems like A.I. is very quickly becoming a marketing term that just means “computers doing computer stuff.” 

Marketing terms are fine, but A.I. is basically the tech equivalent of “organic.” Companies are racing to sprinkle it on every product because it’s the current trend and no one wants to be left out.

By the way, when I call it a trend, it’s not that I think A.I. as a technology is going away. I just don’t think it’s here yet. 

For example, ChatGPT–which is arguably the most popular “A.I.” product–isn’t artificial intelligence, even though it’s made by a company called OpenAI. It’s very cool, yes, but it isn’t a computer thinking for itself. It’s just a model that parses your prompt, and then decides what the most likely combination of words should be based on the massive set of data that humans used to train it. That’s not to say it isn’t impressive. It’s just not A.I.

Here’s another way of looking at it. What if you have a bunch of photos and you ask a piece of software to organize them according to the date and time they were taken–is that artificial intelligence? No, it’s just a computer doing computer stuff. It just looks at the metadata and organizes it based on something a human programmed.

There was a time when it was cutting-edge impressive technology. Today, it’s so commonplace we just expect photo software to do it whether we ask or not.

Now, what if that software is able to organize photos based on who is in them, is that A.I.? No. It’s still just a computer doing computer stuff. It’s very cool, and looking at pixels in an image to match them to a computerized map of other pixels is definitely more complex than looking at dates, but it’s still just organizing files based on something a human programmed.

What if that software could remove a distracting element from the background of one of those photos? Humans can do that using software like Photoshop or Pixelmator Pro. You can use the various brushes and tools to manually paint new pixels that look like the background over the old ones of whatever you don’t want in your picture. 

Some software can do it automatically. Google Photos has been able to do that on its own with what it calls Magic Eraser using A.I., the company says. Now, the company says it can even edit your photos, including by reframing your subject and painting in whatever is missing. 

But your computer doesn’t know what a good photo is. It only knows what humans have told it is desirable, and it just cuts out the middle steps of dragging a mouse around an interface to give you what you want better and faster than you could do it on your own. That, by the way, has always been the point of using computers. You no longer have to learn Photoshop. Just use Google Photos. It’s just computers doing computer stuff.

Or take Apple. On its recent earnings call, Tim Cook said the company has been building A.I. into features like ECG, fall detection, and crash detection. Is it though? Is that A.I.?

Sure, the company trained its algorithm on massive sets of data to teach your Apple Watch to identify when you’ve fallen off a ladder, but machine learning (ML) isn’t the same as what people think of when you say A.I.

At Google I/O, Rick Osterloh, who heads up the company’s hardware division, said when introducing the new Pixel 7a, that the Pixel has always been a “smartphone built on A.I.” Really? What does that even mean? 

I think it just means that the people at Google have built all sorts of helpful things into the Pixel. It can detect spam calls and messages, edit your photos, and will even call and wait on hold on your behalf. 

The reason all of this matters is that words matter. What you call something matters. It probably doesn’t seem like it because most people don’t understand how computers and software work. You can get away with calling cool features A.I. because we’ve never seen computers that can think for themselves. Instead, we’ll just settle for computers that will have conversations and make things up.

I believe there absolutely will come a day when A.I. will be a thing. We’re just not there yet, and calling cool software features A.I. just because it’s what everyone is talking about does a disservice to your users and to the technology. For now, let’s just stick to what is really is: computers doing computer stuff.