You Agreed to All of It

I want to make an admission that will sound strange coming from someone who has spent twenty years reading incident reports, audit findings, and other people's fine print for a living: I do not read terms of service agreements. Almost nobody does. I recognize the shape of these documents well enough, from professional exposure, that I could probably summarize most of them without opening them at all — which is its own kind of problem, because recognizing a document's shape is not the same as knowing what's actually in the one sitting in front of me. And yet I click "I agree" anyway, virtually every time, on virtually everything. So do you. So does almost everyone reading this.

That fact has been sitting underneath every article in this series so far. Windows, Google, Facebook, Apple, TikTok, LinkedIn, the data brokers you'd never heard of before Article 6 — every single one of them, at some point, can point to a document you agreed to. A checkbox you ticked. A button you clicked. And in a narrow legal sense, they're right. You did agree. This article is about what that agreement actually was, and whether the word "agreement" means anything close to what most people assume it means when a corporation uses it.

A Document Built to Not Be Read

Researchers have tried, more than once, to figure out what it would actually cost a person to read every privacy policy and terms of service agreement they encounter in an ordinary year of using the internet. One widely cited academic estimate put the number at something in the neighborhood of two hundred hours annually — roughly a month of full-time work, spent reading legal documents, just to keep pace with the digital life most people already have. Nobody does this. Nobody could do this and also hold a job.

That gap between what "informed consent" would require and what any human being could plausibly do is not an oversight. It's not a byproduct of lawyers being long-winded because that's what lawyers are like. The length and density of these documents serve a function, and the function is not communication. A document nobody can realistically read cannot realistically be objected to. Friction, in this context, isn't a flaw in the system. It's a feature that happens to look like a flaw.

I don't think most of the people who draft these agreements sit down with the explicit goal of making sure nobody reads them. I think, more realistically, that the incentive to write a document covering every possible legal contingency, combined with the absence of any incentive to write one a normal person would actually finish, produces exactly this outcome without anyone needing to intend it directly. That's a pattern this series keeps returning to. Nobody has to build the system to prefer opacity. They build it to prefer legal coverage, and opacity is just what that kind of legal coverage looks like.

The Click That Counts as a Signature

Here's the part that surprised me most when I actually looked into how this works legally, because I'd always assumed, in the same vague, unexamined way most people assume things, that clicking "I agree" on a wall of text was a somewhat weaker form of consent than signing a paper contract. Legally, in the United States, it generally isn't. Courts have repeatedly upheld the idea that clicking a button next to a link to a lengthy document constitutes acceptance of everything in that document, whether or not you opened the link, whether or not you would have understood it if you had, and whether or not you had any real alternative to clicking it if you wanted to use the product at all.

Some versions of this are even looser than a click. Plenty of services rely on the idea that simply continuing to use something after being told, somewhere, that terms have changed counts as acceptance of the new terms. You don't have to click anything. You don't even have to see the notice, provided the company can show it was reasonably available to you. Keep using the app, and the law generally treats that as your signature.

I don't say this to argue that these agreements are secretly illegal, or that a clever lawyer could get you out of all of them. They're not, and one probably can't. The legal architecture behind this is well established, has survived plenty of court challenges, and isn't going anywhere on its own. What I'm pointing at is narrower and, I think, more useful: the word "agreement" is doing an enormous amount of work in a place where almost none of the things people usually associate with agreeing — reading, understanding, having a real alternative — are actually present.

The Design Choices Nobody Notices Making

Terms of service are the blunt instrument. The more interesting, and in some ways more troubling, layer sits one step down, in the actual design of the screens you interact with every day. Researchers and regulators have a name for this now: dark patterns. Interface choices engineered, deliberately, to push you toward the option the company wants and away from the option you might actually prefer, without ever technically lying to you or removing your ability to choose.

You've encountered these more times than you could count. Most of them were individually too small to notice, which is sort of the point. The box that comes pre-checked to opt you into marketing email, sitting below a dozen other form fields you're focused on filling out correctly. The account you can create in ninety seconds and cancel only by calling a phone number during limited business hours and sitting on hold. The "no thanks, I don't want to save money" link, in gray six-point text, sitting beneath a brightly colored "yes, sign me up" button twelve times its size. Individually, each of these looks almost too petty to be worth writing about. Collected across the dozens of services an average person touches in a week, they add up to a systematic tilt in every decision toward the outcome that benefits the company, engineered by people whose job is specifically to understand how human attention and decision-making work well enough to bend them slightly, consistently, in one direction.

What strikes me about this list, looking at it as a whole, is how little any single item resembles what most people picture when they hear the word "manipulation." There's no lie anywhere in it. Nobody's information is technically false. Every one of these patterns works by using genuine, well-studied facts about how human attention and decision fatigue actually function, and pointing that knowledge at you, thousands of times a day, across every service you use. It's the same asymmetry I described with Facebook's engagement algorithm back in Article 4 — nobody has to design a system to trick you. They design it to convert attention and inertia into an outcome, and the outcome is simply whichever behavior the metric happens to reward.

Some Ground Has Started to Shift

I don't want this to give the impression nothing has ever pushed back against any of this as that wouldn't be honest. Regulators on both sides of the Atlantic have started treating dark patterns as their own category of legal problem, separate from whether the underlying data collection is disclosed anywhere at all. The Federal Trade Commission has brought and settled cases specifically over subscription cancellation flows designed to be harder than the sign-up flow that preceded them. Europe's data protection framework requires, at least on paper, that consent be freely given, specific, and as easy to withdraw as it was to give — a standard that would invalidate a meaningful share of the patterns described above if it were consistently enforced everywhere, which it isn't.

California's approach, which I touched on with the data broker registry in the last article, is part of the same broader shift: a slow, uneven, jurisdiction-by-jurisdiction recognition that "you clicked a button" is not the same thing as "you understood and chose this." None of it amounts to a comprehensive fix. The United States still has no single federal privacy law that would apply this standard everywhere, to everyone, uniformly. What exists is a patchwork, stronger in some places than others, generally trailing years behind the pace at which new patterns get invented. But a patchwork that didn't exist a decade ago now does, and I think that's worth highlighting rather than dismissing as insufficient and leaving it there.

What "You Agreed" Actually Means

So here's where I land, after six articles establishing what gets collected and one article establishing how the collection gets legally authorized: the sentence "you agreed to this" is technically true and practically almost meaningless, and both halves of that matter equally. Technically true, because a court will generally hold you to what you clicked. Practically almost meaningless, because the conditions that would make an agreement actually informed — a document you could realistically read, a choice you could realistically decline without losing access to something you need, language written to be understood rather than merely defensible — were absent in nearly every instance across this entire series.

I keep coming back to something my wife said, which I mentioned all the way back in Article 1, about a relative refusing medicine over not knowing what's in it: "I've seen you eat hot dogs, so that argument is kind of flat." I've used that line as a way of acknowledging something true and slightly uncomfortable about human nature — that nobody reads every ingredient list, nobody fully audits every system they participate in, and that's not a character flaw. It's just what being a person in a complicated world looks like.

But I want to be precise about where that comparison holds and where it stops holding, because I think collapsing the two things together is exactly the mistake this whole series has been trying to avoid. Hot dog ingredients are regulated. They're listed, in a standardized format, by law, and if a company lied about what was in one, there's a real regulatory body that would come after them for it. You can walk into that decision not having read the label, and still be operating inside a system built, at least in principle, to protect you if you had. Nobody engineered the ingredient list to be two hundred hours long. Nobody put "yes, I consent to this" in bold and "no" in six-point gray text underneath it. The asymmetry that exists in a hot dog package is incidental. The asymmetry in a terms of service agreement is the design.

That's the actual difference, and it's the one I'd ask you to consider as this first half of the series closes. Not "did you technically click yes," because of course you did, over and over, on nearly everything, the same as I have. The better question is whether that "yes" was ever built to be understood, or whether it was built to be clicked past. Once you can see that difference clearly, in one company's design after another, it becomes very hard to keep calling the whole arrangement "consent" without putting the word in quotation marks (and maybe rolling your eyes at the same time).

I'm not writing any of this to make you feel foolish for every account you've ever created, every trial you forgot to cancel, every box you left checked because you didn't notice it was there. I've done all of the same things, plenty of times, despite two decades of professional reasons to know better. The point was never to shame anyone into reading two hundred hours of legal text a year. Nobody's going to do that, including me. The point is narrower: once you can see the mechanism clearly — the length, the defaults, the friction built asymmetrically in one direction — you get to decide what to do about it with your eyes open, rather than carrying around a vague, unexamined assumption that somewhere along the way, you actually understood and agreed to all of this.

The question of what you do with this information, now that you can see it clearly, is where Act I of this series has been heading from the very first article. It's also where the series itself has to make a choice about what kind of answer it's going to offer. In the next article, I want to talk honestly about the two different paths people tend to take once they get here, why I don't think either one is more legitimate than the other, and which one this series is going to follow the rest of the way.