Big Tech and Consent

29 Jan 2025 by cthos
1076 words

_{A garbage fire seemed like an appropriate cover image. konstantinos69 @ Shutterstock #1442040206}

This is going to be a rant. It'll hopefully be very short, but no promises. Since LLMs became a thing it feels like all the firms that are shoving them into everything have completely given up on the veneer that “asking the customer if they want a thing” is something they should even consider. I'm not super surprised by this, but I am horribly disappointed.

Let's start by getting this out of the way. As of 18.3 Apple Intelligence is now on by default. You can't disable it during onboarding, it'll download a 7 GB model file to your computer or device, and then you have to opt out of it later after it's done its installation. I am displeased by this decision. They could have included a toggle during onboarding, but they instead are forcing a download on you and then clearing up space later.

This screams “not enough people are opting into this thing so we need to mandate it to pump our installation stats”. They're either hoping people just haven't experienced the “joy” of Apple intelligence and will be convinced when it's foisted upon them. Either that or some manager somewhere has an installations OKR and the quickest way to hit that is to make a feature opt-out. The fact that it's not even a choice before a multi-gig download happens (near as I can tell, I'm not installing this stuff), is especially egregious.

Microsoft is doing the same thing. First they forced a Copilot logo onto the taskbar, next they're putting their spyware recall feature on as many computers as they can, and they've added copilot to Office 365 subscriptions and making it very difficult to opt out of paying more for the privilege.

Google's doing the same thing. They've added Gemini to Workspace and are also raising prices.

You will accept the LLM and you will like it.

So, the main reason you include a feature as opt-in, especially when not giving the user an option to proactively consent, is because you don't think people are going to turn that option on. The reason you bundle a “feature” with software and jack the prices up is when you can't make a profit on the product by itself.

They didn't even bother to do the illusion of choice that I've seen apps include, where there's a prompt but the “accept button” is a bright color, and you're hoping people will click through without reading. It's just on. It's on you to go out of your way to turn the feature off later.

There are very few reasons why features should be opt-out, and almost no reasons why they should silently opt you in.

Look, the legitimate reason for having an opt-out is for cases where the default option is going to be good for the most people, but it has considerations that will require it to be disabled or augmented. For example, in video games the artistic vision might need to include a motion blur, but that causes accessibility problems for a lot of people. That should at minimum be opt-out with a consent prompt. Right at the beginning, “hey this has a lot of motion blur, do you want to turn that off?”. Same deal with color choices, and the like.

Heck, the web even has a browser setting to reduce motion, and it's good form to respect that setting — or ideally not even including a lot of motion in the first place.

Other good opt-out reasons might be for security. When you install a VPN on GrapeheneOS, for example, it will set that VPN to “always on” and “block traffic when the VPN is off”. That's the safest option, but it also has some very annoying downsides. Either way, it does prompt you to inform you of this behavior.

But, in no case should an opt-out feature cause harm. In Apple intelligence's case, there are so many ethical and data privacy considerations that an opt-out is especially jarring to me. Not only that, it eats out an additional 7 GB of storage space, that you may not even want, including a download for that file. There are still people on bandwidth constrained internet connections, Apple.

The Google and Microsoft examples are less egregious, but their marketing of “it's now included for free” (but ignore the fact that we're raising prices) makes me think they know this is a bad idea that they've tried to disguise behind weasel language.

Regardless, the latest slate of LLM news — with them being shoved everywhere makes me think that big tech is so desperate to get their return on investment on the LLM machine that they've given up all pretext of consent. Let's give some more examples:

• Gemini will send you a text message to tell you that you can message it at any time if you have the Google Messages app installed.
• Microsoft Edge will summon copilot if you so much as hover over the button that you can't get rid of (this is a distinct button from the one they put on the toolbar)
• Windows is showing Full screen ads on Windows 10 PCs
• Windows is showing Ads in the start menu
• Meta let chatbots loose in group chats
• Meta's making AI Influencers
• Meta “accidentally” blocks mentions of Linux

None of these things show any modicum of respect for their users. The Enshittification continues unabated.

I'm especially disappointed in Apple. They're more slowly running down the garbage train and disrespect for their customer than the other tech firms, but they're still slowly rolling down the same hill.

Microsoft has gotten so bad I've totally sworn off Windows on my personal devices. Their reach is invasive, and they feel like you're beholden to them. Fortunately Gaming on Linux is better than ever, and I'm not missing a Windows gaming PC.

Anyway, it feels like we're accelerating away from tech making actual innovations, and into an era where (US based, at least) tech companies are just going to force-feed us slop. I don't like it.

AI tech rants