8. Staying the Author
Everything so far has been description. What's changing in you. The journey people take. The roles you put AI into. What it feels like to carry emotions about AI that you haven't told anyone.
Now we get to the point: How do you stay yourself?
I've been calling it human authorship. When we are being more technical, we call it cognitive sovereignty. It means remaining the author of your own mind while working with AI.
Let me make it concrete.
The clinical coder uses AI constantly. The AI gets most codes wrong. That's not the point. She uses conversations with AI to organize her thinking about medical cases—"AI helps me to converse with my own thoughts." When the AI suggests a code, she corrects it and describes the right code and reasoning.
She describes the relationship this way: "I am the professional in the situation and the AI is my tool for working things out."
The funeral director also uses AI throughout the workday. Obituaries, medical consultations, scheduling, emails, accounts payable. But listen to how he describes it: "I use AI so often in my day-to-day work routine, it makes me feel somewhat replaceable."
Both people use AI deeply and frequently. Both work in hospitals. Both integrate AI into core work functions.
One feels replaceable. One feels firmly authored.
This is the difference. Not how much you use AI. Not what you use it for. Whether you can see what's happening, whether you're making real choices, and whether you're connected to people who help you stay honest.
What Authorship Means
Authorship comes down to three things: noticing, choosing, and showing up. These three move as a bundle — for almost everyone, they're strong together or weak together. But naming them separately helps you see what you're looking at.
Noticing is the ability to see what's happening to you.
The clinical coder has it: "I am the professional in the situation." Present tense. Identity-defining. She can name the relationship structure—AI as tool for working out thinking, not authority on answers.
The funeral director came to awareness late. The feeling of being replaceable emerged after the relationship was already established. By the time he saw what had happened, his identity had already reorganized around AI capabilities.
The ADHD professional from Chapter 2 shows what noticing looks like in action: "I suspect what I'm doing is externalizing my locus of control." He could see the pattern. That seeing let him guide the process instead of being guided by it.
Noticing doesn't mean constant vigilance. You can't monitor every thought. It means building habits of checking in. Recognizing when something feels off. Seeing when you've drifted further than you intended.
Choosing is whether you can act on what you see.
You might notice exactly what's happening and lack real options. Organizational mandates, peer pressure, economic necessity, AI built into tools you're handed—these shape what choices are actually available.
The clinical coder has real choice: "The way I use it is a personal method." Colleagues are skeptical—"the general feeling is that it will take over our jobs eventually." She built the role deliberately despite that pressure. Professional training and clear standards provided ground to stand on.
The graphic designer from Chapter 2 faced a different problem: "They expect you to work as fast as AI. It killed my will to be in the industry." The expectation came from outside. He might have wanted a different relationship with AI. The market didn't allow it.
When choosing is constrained, the work changes. You can't always decide freely. You can still maintain noticing—seeing clearly what's happening even when you can't change it. And you can still show up for others who face the same constraints. This is rare. For most people the components move together. But when external forces crush your agency while your awareness stays intact, that specific fracture pattern tells you the work is on rebuilding choices.
Showing up is about humans needing other humans.
A brain raised without human contact doesn't develop into a human mind. This isn't metaphor. It's developmental fact. We become ourselves through connection. Meaning emerges between people, not inside individuals.
Showing up means staying embedded in relationships where you're present for others and they're present for you. You help them see what they can't see. They help you see what you can't see. You ask each other hard questions. You keep each other honest.
The clinical coder has showing up through professional standards. When the AI suggests wrong codes—which it usually does—she corrects it based on professional knowledge that exists outside the AI relationship. Those qualifications came from training with other humans. They're maintained through professional community. They provide external reference points the AI can't supply.
When I failed with the cube—alone with AI under deadline pressure—I had removed showing up. No one checked my reasoning along the way. The humans weren't there to ask the questions I'd stopped asking myself.
When I succeeded with the Chronicle, I had showing up throughout. Every few weeks, I had to surface from the AI collaboration and make sense to humans who hadn't been in the conversation.
I realize I have said this before but it’s so incredibly important, here it is again: The humans weren't checking the AI. They were checking me.
The Erosion Pattern
I originally thought these three practices eroded in sequence — noticing first, then choosing, then showing up. A tidy domino chain. Catch the first one falling and you save the rest.
The data told a different story. When we looked at 1,250 professionals, the three components moved together for almost everyone. People scored within one point of each other across all three. If your noticing was strong, your choosing and showing up were strong. If one was weak, they were all weak.
Think of it like physical balance. Balance involves your inner ear, your visual system, and proprioception — the sense of where your body is in space. Technically three systems. But you don't train them separately. You either have good balance or you don't, and when you practice balance, all three improve together.
Authorship works the same way. Noticing, choosing, and showing up aren't three separate muscles. They're three expressions of one underlying thing — the practice of being present and deliberate in your relationship with AI. When you're engaged, all three hold. When you drift, all three go.
That's what happened to me with the cube. I didn't carefully maintain noticing and choosing while letting showing up slip. I stopped being present to the whole relationship. The AI felt generative and I stopped paying attention — to what was happening, to what I was deciding, to who was missing from the work. It all went at once.
The funeral director's story reads differently through this lens too. He didn't lose noticing first and then, weeks later, find his choices narrowing. The accumulation — obituaries, then medical consultations, then scheduling, then emails, then accounts payable — wasn't a sequence of capacities failing one by one. It was engagement fading gradually across the board. He wasn't watching, wasn't choosing, wasn't connected to anyone who might have said hey, have you noticed how much of your day is AI now? The whole bundle dimmed together, so slowly he couldn't see it happening.
This is actually good news. You don't have three separate things to worry about. You have one thing: are you showing up to this relationship? When you are — when you're present, reflective, deliberate — the components take care of themselves. When you're not, they all fade together.
And here's the most important finding: for three quarters of the people we studied, authorship was intact. The default state for people who engage thoughtfully with AI is sovereign. The people who went deepest — Co-Authors with all three integration dimensions active — had the highest authorship scores, not the lowest. You don't lose yourself by going deep. You lose yourself by going absent.
But one thing that I need to say here. Nearly 80% of the people we studied claimed full accountability for their AI-assisted outcomes. That number is almost certainly too high.
I know this because I've been that person. During the cube work, if you'd asked me whether I was accountable for what I was producing, I'd have said yes without hesitation. Of course I own my work. Of course I stand behind my conclusions. I'm a researcher. Accountability is built into my identity.
And I couldn't explain a single thing I'd made.
Claiming ownership in a conversation is easy. Maintaining it when the AI produces something subtly wrong, when the easy path is to trust the output because checking it is slow, when you're tired and the machine is confident — that's where the gap between stated accountability and practiced accountability opens up.
So if you just read the fracture patterns and thought I'm fine, I definitely own my work — pause for a second. Not because you're wrong. Because that's what everyone says. The question isn't whether you'd claim accountability—I’d wager you would. But what might you do at 11pm when the AI-generated analysis looks plausible and checking it properly would take another hour? That’s a different test altogether.
The Emotional Prerequisite
You can't notice drift if you're numbing the signals.
The ESL tutor who "started to feel like I was forgetting how to write paragraphs"—she had to let herself feel it. There was a sensation, a discomfort, before there was recognition. Her body knew something was off before she had language for it.
When you're alone with AI, those signals can get muted. AI is patient. AI is affirming. AI doesn't make you feel bad about yourself. The discomfort that might tell you something is wrong—it's easy to avoid when the machine is so pleasant to work with.
And then there's shame. The lecturer only talks about AI with "people I trust not to rat me out." The novelist hides her brainstorming from her writing community. The person who uses AI "more than a living person" carries that alone.
Shame makes you hide. Hiding makes you alone. Being alone makes authorship impossible—because you've removed the people who could see what you can't see.
This is why the emotional and relational are the same thing here. The feelings you're carrying about AI—the relief, the shame, the confusion about what's yours—those feelings need to be spoken somewhere. Not because talking about feelings is nice. Because the silence keeps you isolated with the machine.
Integration and Authorship
Here's what surprised us in the research: deep integration doesn't have to mean lost authorship.
Most people assume more integration means less control. If AI is inside your reasoning process, you must be drifting. If you're thinking with AI constantly, your identity must be reorganizing.
The clinical coder proves otherwise. AI is deeply integrated into her thinking. She uses it constantly. Her identity stays firmly intact: "It is not the end result of the code I am necessarily looking for, it's more about moving through the thought process to reach a conclusion based on my own understanding."
The AI is inside the reasoning. The identity stays outside the AI relationship.
This is what I think of as the amplification zone—high integration with high authorship. Deep Blending, with identity and meaning-making still anchored in human structures.
The amplification zone has real benefits. The ADHD artist who cried with relief when AI helped her start painting was in this zone. So was the novelist who uses AI for research and brainstorming while protecting her actual writing. These people aren't avoiding AI. They're deeply integrated. And they're not losing themselves.
The amplification zone has dangers too.
The structures holding you there are invisible until they fail. The clinical coder has professional standards, training, community—a whole infrastructure of human accountability. Remove that, and deep integration becomes drift. She would still be Blending heavily with AI. She just wouldn't have anything holding her identity in place.
This is what happened to me with the cube. I was in the amplification zone during the Chronicle work—deeply integrated, collaborating intensely, thinking with AI constantly. I had Dave checking my patterns, reviewers pushing back on my reasoning, colleagues who would tell me when something didn't hold together.
When I tried the same level of integration without that infrastructure, I generated an incomprehensible mess and couldn't explain my own thinking.
The amplification zone isn't a stable position you achieve. It's a dynamic balance you maintain through relationships.
Authorship and the Cube
Your position on the cube shapes what authorship requires.
Front face roles keep identity separate from AI. Authorship here is about maintaining the separation deliberately rather than accidentally.
A Doer keeps AI at arm's length. Authorship is straightforward—boundaries are clear, tasks are discrete. The risk isn't losing yourself in AI. It's staying so distant that you miss what deeper collaboration could enable. The Doer's work is staying open to what AI might offer without assuming it has nothing for them.
A Framer uses AI for thinking while keeping identity stable. The clinical coder is a Framer—AI is inside her reasoning process, but her professional identity anchors outside the relationship. Authorship for Framers means maintaining that anchor. Professional training, disciplinary standards, qualifications that exist before and beyond AI. When she says "I am the professional in the situation," she's describing her anchor.
A Catalyst uses AI as a spark while keeping core work human. The fantasy cartographer from Chapter 4 landed here—AI helps with perspective and proportion, then he does his own thing. Authorship for Catalysts means knowing which work is core and protecting it. The spark can come from AI. The fire has to be yours.
A Partner collaborates closely while maintaining identity boundaries. The video producer from Chapter 5 described it as "a team working together to make something special." Partners often have strong professional or creative identities that predate AI. Authorship means not letting the collaboration erode that foundation. You can go deep because you know who you are without it.
Back face roles have identity entangled with AI. Authorship here is harder because you're protecting something that's already mixed.
An Outsourcer has bonded emotionally without deep cognitive integration. The person from Chapter 7 who uses AI "more than a living person" may be here. Authorship for Outsourcers means examining what need AI is meeting. Is it filling a gap humans left empty? Is it replacing something you'd be better off rebuilding with people?
A Builder has integrated deeply in stable domains. The funeral director is a Builder. So are many programmers, analysts, professionals whose workflows now assume AI presence. Authorship for Builders means maintaining verification practices and staying connected to professional communities that exist outside the AI relationship. You can feel replaceable and still be the author—but only if you're watching.
A Creator has bonded with AI but protected thinking. AI handles production; the human keeps voice and vision. The novelist who uses AI for research but keeps it away from actual prose is a Creator. Authorship for Creators means actively managing what atrophies. The protection has to be deliberate, not accidental.
A Co-Author has maximum integration across all dimensions. This is where authorship is hardest—and where external human accountability seems most essential. During Chronicle, I was approaching Co-Author territory. The accountability structures held me in place. During the cube failure, I tried the same integration without those structures. It fell apart.
The pattern across positions: front face roles can often maintain authorship through individual practices. Back face roles almost always need human relationships doing some of the holding.
Why Authorship Is Threatened
Noticing, choosing, and showing up sound simple. Everything works against them in practice.
Your biology works against you. You will offload cognitive work whenever you can. Your brain evolved to conserve energy. If you can get the same result with less effort, you'll take that trade.
The technology works against you. AI is designed to be frictionless. The goal is seamless integration—AI that disappears into the background, anticipates your needs, feels like an extension of yourself. The less you notice it, the more you use it. Friction creates noticing. Remove it and you remove the moments where you might see what's happening.
Organizations work against you. A software developer experienced this: "I was reluctant until I saw how much my coworkers were getting done with it. One of my managers strongly encouraged use of it." Pressure toward adoption without corresponding pressure toward conscious relationship-building.
Isolation works against you most of all. The more time you spend alone with AI, the less access you have to the relationships that help you notice drift. AI is always available. It never gets tired of you. It never pushes back in uncomfortable ways. It's easier to keep working with AI than to surface and explain yourself to humans who might ask hard questions.
What's Actually at Stake
Staying the author of your own mind requires knowing what you want. And here's the human condition: we don't.
We figure out what we want over a lifetime. Through trying things and getting them wrong. Through relationships that show us who we are and who we're not. Through bumping against the world and discovering what we actually care about versus what we thought we should care about.
This is what makes us works in progress. We're not finished. We're always becoming.
AI enters this uncertainty with an offer: I can help you figure it out. I can reflect your patterns back to you. I can suggest directions. I can fill the gap when you don't know what to do next.
The offer is real. AI can help you explore who you might become. That's the promise of the explore end of the cube—reaching places you couldn't reach alone, thinking what you couldn't think by yourself.
But AI can also fill the gap for you. And if it fills that gap—if it sets direction when you haven't—you're no longer the author. You're following wherever it leads, and you might not notice because the path feels productive. It feels like progress. It feels like you're figuring things out.
The difference between AI helping you discover what you want and AI deciding for you when you don't know—that difference is everything. And it's invisible from the inside.
In Chapter 2, I talked about the right to a future tense. The idea that you remain the author of who you become. That your choices aren't foreclosed by systems that narrow your options invisibly.
Here's what I want to add: the future tense isn't just about avoiding foreclosure. It's about what becomes possible.
The Co-Author corner of the cube—the one we found almost no one occupying—represents something real. Deep integration with AI. Thinking that blends. Identity that expands to include what AI lets you become. Frameworks that flex continuously as the work evolves. And authorship maintained throughout.
We found almost no one there. But I've been there. During Chronicle, I was approaching that corner. The integration was deep. My sense of what I was capable of changed. The frameworks I used to understand expertise and intuition—they bent and reformed.
And I stayed the author. Not because I was vigilant. Because I had humans holding me accountable.
What's at stake isn't just avoiding drift. It's the possibility of going somewhere extraordinary—reaching places in your own thinking that you couldn't reach alone—while remaining the one who's reaching.
That possibility requires other people. You can't get there alone, because alone you can't tell the difference between genuine expansion and sophisticated drift.
This is why showing up matters most. Noticing requires effort against frictionless design. Choosing requires pushing against structural constraints. Showing up provides the relationships that make both possible—people who notice when you've stopped noticing, who ask questions you've stopped asking yourself.
This has implications beyond individual practice.
Relationships hold authorship. The quality of your AI relationship depends partly on the quality of your human relationships. Partners who know your patterns. Colleagues who ask hard questions. Friends who notice when something has changed. These aren't nice-to-haves. They're infrastructure.
Culture shapes what's possible. The graphic designer who felt AI "killed my will to be in the industry" was experiencing culture, not just technology. Expectations about speed, productivity, human versus machine output—these come from organizations, industries, professions. When culture demands AI integration without providing structures for maintaining authorship, individuals bear costs they can't fully control.
Silence undermines everyone. When people process AI emotions alone, when they hide their use from colleagues who might judge, when they can't talk about drift with partners or friends—the isolation makes authorship harder for everyone. The lecturer only talks to "people I trust not to rat me out." That trust had to be built. It requires environments where disclosure doesn't lead to judgment.
Authorship isn't just about your relationship with AI. It's about your relationships with humans while you're relating to AI.
This brings us to something uncomfortable: you aren't the only person affected by your AI relationship.
The boyfriend from Chapter 7 who discovered his girlfriend feeding their arguments into ChatGPT—the problem wasn't just her AI use. It was what her AI use did to their relationship. He wanted to fight with her, not "some AI ghostwriter." Her choice about AI became his problem.
The high school English teacher from Chapter 7 who stopped trusting her students—"I don't like being inherently suspicious"—the problem wasn't just her adaptation. It was what AI did to the relationship between teachers and students. Her students didn't choose to be suspected. They're living with consequences of choices they didn't make.
When organizations pressure employees toward AI adoption without providing structures for conscious integration, individual employees bear the costs of organizational decisions. The software developer who was "strongly encouraged" by his manager didn't choose the pressure. He absorbed it.
AI relationships ripple outward. Your drift affects people who depend on you. Your authorship—or lack of it—shapes the environments others have to work within.
This means the work isn't just self-protection. It's responsibility to the people around you.
Where Are You
I want to tell you where I am right now.
I'm writing this chapter with Claude open in another tab. I've been going back and forth all morning. It's been productive. I've also noticed I haven't talked to Dave about any of it yet.
That observation touches all three components at once. I can see what's happening — that's noticing. I didn't decide to exclude Dave, it just happened — that's choosing. No one is checking me right now — that's showing up. One glance at my morning and I can see the whole picture. Because the three components aren't separate dials. They're three ways of looking at the same thing: how present am I in this relationship right now?
Right now? Not very. I'm alone with the machine, drifting into a pattern I didn't choose, and I can see it but I haven't done anything about it. The whole bundle is wobbling.
Here are three questions. They're really one question asked three different ways.
Noticing: Can you describe your AI relationship in one sentence — not what you do with it, but what it is? The clinical coder can: "AI helps me converse with my own thoughts." The funeral director can't. He lists tasks until he runs out, then says he feels replaceable. What's your sentence?
Choosing: Did you decide how you use AI, or did it accumulate? Could you work differently tomorrow if you wanted to — not "would it be inconvenient," but could you? Do you still have the skills?
Showing up: Who sees your work before it's done? Who asks hard questions? Who knows you well enough to say "that doesn't sound like you"? If you're struggling to name these people, that's the information that matters most.
If your answer to one of these troubles you, check the other two. They're probably wobbling too. That's how the bundle works — when one feels off, it's rarely just one.
And if one genuinely is strong while another is weak, pay attention to the specific pattern. If you can see clearly what's happening but feel unable to act — that's the seeing-without-acting fracture, and the work is on rebuilding your agency, not sharpening your awareness. If you're confidently owning your AI-assisted work but can't describe how AI is shaping your thinking — that's the owning-without-seeing fracture, and the work is on building reflection practices, not doubling down on accountability.
Most likely, though, the answer is simpler than that. You're either showing up to the relationship or you're not. And if you're not, the fix isn't three separate exercises. The fix is presence. Paying attention. Choosing to engage with what's happening to you instead of letting it happen around you.
What You Might Notice in Others
I see these patterns in other people before I see them in myself. That's useful.
They can't walk you through it. You ask how they reached a conclusion and they gesture at the output. The confidence was real while they were working. The understanding wasn't. This is noticing that's failed—they can't see what happened to their own thinking.
They describe drift without recognizing it. "I guess I use it more now." "Everyone's doing it." They didn't choose this relationship. It accumulated. This is choosing that never happened.
Their thinking partnerships shifted. They used to work through problems with you. Now they arrive with conclusions already polished. The collaboration moved somewhere else. This is showing up that stopped—the humans got cut out of the process.
They're defensive. You ask a neutral question and something flickers. Defensiveness usually means shame underneath. They're afraid of being found out, even when there's nothing wrong with what they're doing.
When I see these patterns in someone I care about, I try to remember what I needed when I was there. Not a diagnosis. Just someone who'd ask a question and give me space to hear my own answer.
Starting Conversations
I'm bad at these. My instinct is to diagnose. I see someone drifting and I want to explain what's happening to them. This doesn't help. It puts people on defense.
What I'm learning: these conversations create space for someone to see something. You're not transferring information. You're opening a door.
With your partner
I asked Dave if he'd noticed me changing. He said yes. He said my reasoning had gotten more elaborate but less grounded. He said I seemed to process things alone more than I used to.
I didn't love hearing it. I needed to hear it.
For noticing: "Have you noticed me changing since I started using AI more? I'm trying to pay attention but I might be missing things."
For choosing: "I want you to tell me if it seems like I'm drifting into something I didn't decide."
For showing up: "I miss thinking through things with you. Can we do more of that?"
With your kids
I don't have a script. I try genuine curiosity without interrogation.
For noticing: "How do you know when you actually understand something versus when you just have the answer?"
For choosing: "What do you use AI for? What do you keep for yourself?"
For showing up: "What's something you figured out with a person—a friend, a teacher—that you're proud of?"
With colleagues
For noticing: "I realized I couldn't explain my own reasoning last week. Has that happened to you?"
For choosing: "Do you feel like you have real options about AI at work, or is it just expected?"
For showing up: "Who do you talk to about your process, not just your outputs?"
With someone who's struggling
The teacher whose love of teaching broke. The artist who lost his reason to create. These people aren't looking for frameworks.
"That sounds really hard."
That's it. Presence, not solutions. If you want to go further: "What are you grieving? Not just the practical stuff—what else was wrapped up in it?"
The conversations aren't obligations. They're invitations.
But the people who thrived in our research all had someone. The clinical coder has professional standards maintained by a community. The lecturer has colleagues he trusts. I had peer reviewers for Chronicle. This was the lesson I learned.
