“AI features” are the new bloatware

It seems that, all of a sudden, all of our gadgets have found some way to include “AI” features of some kind. I’ve even seen “AI” screen protectors, whatever that may be.

The term “AI” has become very popular, but the features it’s attached to are nothing new.

We’ve been here before

Now, I’m not some sort of Luddite that doesn’t understand how cutting-edge technologies like LLMs, vision systems, agentic AI, and all the other technologies that are now mature enough will have a massive impact on the world.

Spotify and ChatGPT logos floating around with a playlist in the background. Credit: Lucas Gouveia/How-To Geek | Spotify

We’re not actually talking about that aspect of the current AI boom. No, one part of it is how companies are trying to apply that technology to every last thing they can think of to see what sticks. LLMs and other machine learning technologies are flexible tools, which means it’s a hammer that makes everything look like a nail. It’s the only thing that explains why Microsoft, for example, thought it would be a good idea to implement a feature where an AI captures and examines your computer screen every few seconds to “help” you. They seemed surprised when people were too creeped out by the idea and they had to backtrack, at least partly.

At least for now.

The other side of this coin is the age-old marketing practice of adopting new buzzwords that the public don’t really understand. That’s why you get “nanotechnology” detergent and “quantum” everything. Even when neither of those terms have any business being near the product in question. That’s the phase of AI we’re in right now.

AI that no one asked for

Copilot+ PC artwork Credit: Microsoft

Every phone, every OS, everything with or without a screen seems to have some sort of app or feature that has “AI” in the name. Photo enhancers, voice note takers, writing assistants, and much more besides. Some of it could be genuinely useful, but in many cases, these features are shoehorned in and don’t add value. They just disrupt the perfectly-adequate workflows people were already using with no issue.

Some of this stuff runs from the cloud, some of it runs locally. It can have an impact on device performance, but the impact on usability is much more of an issue in my opinion. Just like the bloatware we used to know, opting out, removing, or disabling this stuff is intentionally tricky, or just flat-out impossible.

The Illusion of intelligence

Face recognition and personal identification technologies in street surveillance cameras. Credit: Trismegist san / Shutterstock.com

“Artificial Intelligence” is a broad field that encompasses a lot of different technologies, many of which have been around for decades. Various algorithms and other intelligent software and hardware devices are technically AI. AI is broader than machine learning, and some marketers have taken advantage of this by capitalizing on this technicality and the perception of the public that “AI” means something like ChatGPT or other sophisticated technologies related to it.

So face recognition in camera apps or basic predictive text like we’ve always had can look like “new” features, because they get that “new look, same taste” treatment breakfast cereal goes through every few years.

The hidden costs of onboard AI

Apple Intelligence is helping create a playlist in Apple Music. Credit: Nathaniel Pangaro / How-To Geek | Apple

As for the “true” AI stuff that runs locally on your computer. Well, whether you use it or not, you’re paying for it. This isn’t like uninstalling Candy Crush from a new Windows installation. It’s more than the cost of a CoPilot key on your new laptop. That special NPU chip in your computer or phone wasn’t free. So if you don’t benefit from it, it’s to your detriment.

Apart from the fact that most current NPUs are too weak to really run anything AI locally that’s substantial, computer makers can ask for more money or make their computers seem more advanced than they are by calling them “AI” PCs.

A smarter user backlash is coming

Hopefully, this phase won’t last too long. We’re not stupid, so people will learn to take any mention of “AI” with a pinch of salt and an eye-roll. At some point, marketers will move on to some other buzzword (quantum will probably come back) and this issue will solve itself. When it comes to the “real” AI, the high-end stuff we need the cloud for right now, it’s going to be a matter of time.

At some point, the models will be more efficient and the local AI processors will be more powerful. When these meet in the ideal middle, then we should have access to AI features and technology that not only don’t make you dependent on a remote subscription, but can also do things that are genuinely useful, and not a gimmick.


I’ve been annoyed by bloatware for as long as I can remember, but unless we suddenly develop some sort of post-capitalist utopia where there’s no incentive to market, sell, and profit as much as possible, bloatware of some sort will remain a fact of life.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top