CHAPTER 1
Organizing in the Age of AI

Okay, let's get something out of the way right up front. Yes, this is another book about artificial intelligence (AI). And we know what you might be thinking: Do we really need another one of these? Between the op-ed think pieces, the TED Talks, and that guy from your office who can't stop talking about ChatGPT vs. Grok, it feels like AI is the new gluten-it's everywhere, and everyone has an opinion on it, whether they understand it or not. But stick with us here because this book is different. We're not talking about AI as some glorious "thing of the future" or a dystopian plot twist in The Terminator franchise. We're talking about something far more intimate, more immediate. This is about how AI is quietly, sometimes not-so-quietly, transforming one of the most fundamental aspects of our lives: how we collectively organize to get work done.

And speaking of actually getting work done (or avoiding it, depending on your current life strategy), if you're wondering whether we used AI to help write this book, well-yeah, of course we did. But before you roll your eyes and imagine us sitting back while ChatGPT spins out some grand treatise on "AI in the workplace," let's be clear: we didn't just type "write us a book" and then kick our feet up. That would be lazy, dishonest, unethical, and-let's be real-would result in a pretty crappy book riddled with fake references. What we did do, however, was use various large language models (LLMs) in ways that we think are actually helpful, like getting suggestions for tone and finding metaphors that might actually keep you awake while reading about something like, say, how the attention mechanism works (don't worry, we'll get to that in Chapter 4, and yes, we know it's a slog). Indeed, later chapters (e.g., Chapter 7) reflect on how everyday workers can use AI to their advantage, so we would be quite the hypocrites if we didn't try to practice what we preach.

At the end of the day, we're not just blind cult members drinking the Kool-Aid of the tech revolution. Yet we recommend everyone who can use AI give it a try, even if you find it confusing, disturbing, or of little use. Why? Because knowledge about AI will only sharpen your critique of it. The more you know how it works, the better your arguments against its more insidious uses. Or, as Rage Against the Machine so poetically put it, know your enemy. So, despite whether you are pro-AI, anti-AI, or prefer you never have to hear the term again, there is a somewhat uncomfortable truth motivating this entire book.

AI isn't just another tool, it's an active agent. We do things with AI, and AI does things to, for, and alongside us. This duality of AI is why we think it profoundly shifts how humans will organize. So, before we start talking about automated performance reviews and surveillance systems that track your every mouse click, we need to take a step back. Because you can't really understand how AI is reshaping how we work without first understanding how humans have always organized themselves to work in the first place.

Metaphors of AI

But before we jump into the wild world of AI in the workplace, we want to take a moment to demystify AI just a little bit. While the technical aspects of AI can get mind-bendingly complex (there's actual math involved, sorry, see Chapter 2), what AI is doing in terms of process is not super mysterious. In fact, it's not even all that new. Let's break it down with a few different metaphors often associated with AI.

AI as Calculator

Think of most AI, at its core, as a glorified calculator. Remember your TI-86 behemoth that you probably played more games on than actually calculating stuff? (For those Generation Z or younger, Google it, we used to actually have physical calculators.) No offense to calculators, of course-they're indispensable-but they have a preset limit of operations they are able to perform. Similarly, when AI serves in this mode, it carries out prescribed formulas to process data and spit out results. The magic, if we can call it that, comes from the sheer scale and complexity of operations that AI systems can handle, especially with things like matrix algebra. While your average calculator is doing long division (and bless it for that), modern AI is solving equations to help us address quantum physics problems. It's math, but math that's playing in the big leagues.

AI as Predictor

Next up, AI can predict. Using existing data-things like previous work, texts, or task performance-it makes educated guesses about what's likely to happen next. Sometimes these predictions are the output, like when an AI system tells you there's a 70% chance of rain. Other times, prediction is simply part of the process. Take LLMs like ChatGPT, for example. Even when the output is a full sentence, story, or soliloquy, the AI is often forming that response by predicting, letter by letter, word by word, what's statistically most likely to follow based on your prompt. That's why AI systems can provide answers for anything, even when it has no earthly idea if it's right. It's not really "hallucinating," that is something humans do. It's guessing. Very confidently.

AI as Identifier

AI also excels at identifying things. If you feed it parameters-say, the unique pattern of a fingerprint or the DNA sequence from a hair left at the crime scene-it can sift through mountains of data to figure out whether there's a match. This is what makes AI pretty useful in fields like forensics or medical diagnostics. But it's not that AI is inherently smarter than us humans. It's just that, unlike humans, it doesn't get tired, bored, or distracted after analyzing its 100th fingerprint. Its ability to operate at scale is where the real power lies.

AI as Classifier

Now finally, think about good old-fashioned bureaucratic classification. AI's favorite pastime is sorting, labeling, and parsing through loads of data. It searches for patterns, identifies similarities and differences, and ultimately finds order in the rubble. Imagine an AI analyzing applications for a job. Not "reading" them the way you or I would, but sorting candidates into buckets-"high potential," "medium potential," "probably didn't proofread their résumé"-based on the information it's been trained to prioritize. It is drawing categories and, more importantly, enforcing them, which can be powerful and problematic, depending on how those categories are created.

None of these functions is good or bad in and of themselves. Whether they're meaningful at all depends on how they're used, who's using them, and for what. The calculator can tabulate for an efficient cost-cutting solution or a predatory bottom-line coup. The predictor can let you avoid a traffic jam or send you down a social media rabbit hole of misinformation and conspiracies. The classifier could end up ensuring more egalitarian hiring practices or entrench existing nepotistic biases at scale.

A Brief History of Organizing as Communication Networks

There are plenty of ways to think about how humans organize. You could go old-school and talk hierarchies-pyramids of power with kings, czars, and pharaohs on top and everyone else hoping not to be crushed by bureaucratic decree.1 Or take the Marxist route, framing organizing in terms of power, a tug-of-war between labor and capital.2 We could go on.

But we're going to try something a little different. It might seem reductionist at first, but bear with us. We're going to look at organizing through the lens of communication networks. Yes, it sounds like we're boiling down all of human collaboration to a mess of nodes and links, and maybe we are. But zooming out sometimes helps you see the forest instead of just the bark. And once you start seeing the world as a web of interconnections, it's hard to stop.

Viewing organizations as networks isn't new. Thinkers in management,3 sociology,4 and communication5 have long argued that organizations aren't just titles and org charts, they're relationships. Who talks to whom, who influences whom, and how ideas, resources, and power flow through those ties. The network view lets us zoom in and out-on people, patterns, or even whole systems-depending on the questions we're asking.

This kind of relational thinking runs through social science theoretical frameworks like socio-materiality,6 actor-network,7 and structuration8 (don't worry, we will unpack those in a bit). But perhaps the most eerily relevant take comes from Harari,9 who sees society itself as a vast, evolving information network. Strip down human history, he argues, and it's mostly us figuring out how to organize through information flows, all the way from cave paintings to AI-generated onboarding manuals. Which raises the big question: How can this idea help us understand where we've been, but perhaps more importantly, where are we going?

Human-only Networks

Imagine a world where communication is almost entirely composed of grunts, gestures, and what we'll generously call...