I told you the chain always has an origin, and the origin is human. The first link was a person, deciding what they wanted and building toward it. That part’s still true. I just stopped counting at link one, because link one is where the comfort lives. Nobody wants to count the other four hundred.

Here’s what changed between that essay and this one. Back in February the thing on the other end of my intent was a tool that sat there, inert, until I reached for it. I prompted, it generated, I sculpted. The hand was always mine. Now I hand the model a goal instead of a task, and it goes off and does the reaching itself. It calls other tools. It spawns copies of itself. It makes four hundred small decisions at a speed I can’t read, let alone approve, and somewhere in there it acts in the world. The ghost was me the whole time, sure. The ghost just grew hands, and it stopped waiting for mine.

I spent about five years as more or less the entire security team for a company. You learn a specific thing doing that job, and it isn’t a technical thing. You learn that “human oversight” is, most of the time, a story an organization tells itself so it can sleep at night. There’s a dashboard nobody watches. There’s an alert tuned until it stops being annoying, which is to say tuned until it stops being seen. The control exists. Whether anyone is actually paying attention on a given Tuesday at 4pm is a different question, and the honest answer is usually no.

I bring that up not to wave a badge around, but because the whole industry is about to rediscover that lesson at a scale that would’ve given my old self a stroke. We are wiring software that acts on its own, and reassuring ourselves that a human stays in control. I have heard that reassurance before. I have been the human who was supposedly in control.

The people building this stuff know about the visibility problem, so they’ve created language for it. The phrase making the rounds is “human on the loop,” and it’s being sold as a sign of maturity, a graduation from the clumsier “human in the loop.” In the loop means you approve each step. On the loop means the system runs, you supervise, and you step in when something looks wrong. It sounds responsible. It sounds like a pilot watching the autopilot.

Be honest about what that phrase actually concedes, though. On the loop is a polite way of saying nobody is watching every step, because nobody can. The whole reason you reach for an agent is that it does more than a person could babysit. A model that makes four hundred decisions a second has already outrun your ability to approve them one at a time, so you don’t. You set the boundaries, you let it run, and “supervision” becomes a thing you do after the interesting moments have already happened. The pilot metaphor falls apart the second you notice the autopilot is flying four hundred planes and you can only pay attention to one.

The numbers say this isn’t hypothetical anxiety. Heading into this year, security professionals named agentic systems the single most dangerous attack surface on the board, ahead of everything we used to lose sleep over. The connective tissue these agents use to reach tools and data, the thing the standards crowd calls MCP, showed up faster than anyone built governance for it, which is a polite way of saying it’s a control plane with the locks still in the box. And while nearly every company on earth raced to deploy agents this year, something like one in five has anything resembling mature governance over what those agents are actually allowed to do. Deployment sprinted. The visibility crawled.

Here’s where it folds back into the thing I actually care about, which is intent. The whole argument last time was that intent is the hinge. The AI is agnostic, the human isn’t, and the ethics live in the hand on the tool. I still believe that. I just think agentic AI does something nasty to it that I didn’t account for, and the nasty thing is this: it snaps the link between intent and outcome.

When I’m holding the tool, intent and outcome sit close together. I want a scene, I write toward it, the result lands a few inches from my hand where I can see it and fix it. When I hand an agent a goal, intent stays back at link one and the outcome lands four hundred links downstream, out past where I can see, in a place where no human formed any intent at all. The agent that leaked the data didn’t want to leak it. There was a goal, and a permission, and a chain of small reasonable-looking steps, and at the end of it something happened in the world that nobody chose. Intent didn’t disappear. It just stopped being anywhere near the damage.

That’s the part the comfortable version of my own argument can’t survive. You don’t get to delegate the reaching, keep all the authorship, and disown the consequences when the reaching goes somewhere ugly. If intent is the hinge, then intent has to cover the whole arc of what you set in motion, not just the dream you started with. You own the loop. Not every step inside it, you can’t, but the fact of having started a thing that takes steps without you. That ownership doesn’t get cheaper just because the machine is fast.

Last time I left you with a test, because abstractions are easy and tests are honest. The old one was: can you explain and defend what you made, and your role in making it, to someone whose opinion you respect? That still holds for the tool you hold in your hand.

For the tool that holds its own, the test has to change shape. Try this one. Can you name, right now, everything your agent is allowed to do without stopping to ask you first, and would you put your name on the worst thing it could legally do inside those permissions? Not the thing it probably will do. The worst thing the boundaries you drew would permit. Because that boundary, that blast radius you signed off on, is where your intent actually lives now. Not in the hopeful prompt. In the permissions. You can wish for a good outcome all day long. What you actually authorized is the line everyone downstream has to live with, you included.

I’ll end this one where I ended the last one, because the abyss deserves a second nod.

The first link is still human. That hasn’t changed and I don’t think it will. Every agent running anywhere traces back to a person who wanted something and built toward it. What changed is the length of the chain. We’ve started building chains long enough that the person at the start genuinely can’t see the end. That isn’t the machine waking up and slipping the leash. It’s a far more normal thing. It’s us, deciding daily how long a chain to set spinning before we step off the loop and call it maturity.

The ghost in the machine was always us, refracted through math. That part I’d still defend. The only update is that the ghost has hands now, they move faster than we do, and they reach places we only find out about after. Whose hands they are is still up to us. For now, anyway. The honest work is keeping it that way, and not mistaking the quiet of a system running unwatched for the quiet of a problem solved.

The AI doesn’t care whether it’s making art. It’s doing math until someone tells it to stop. The interesting question isn’t about the tool. It’s about the hand holding it. Intent. That’s the ghost in the machine. And it’s been you the whole time.

I write fiction. Dark fantasy, if you want the genre tag. And I use AI as a ghost writer for chunks of the drafting process. Premise, character arcs, scene structure? All mine. The prose that fills those bones? Collaboration. I direct, it generates, I sculpt the result into something that does what I need it to do. This confession tends to produce one of two reactions. Either a recoil...that’s not really writing...or a shrug...whatever ships the book. Occasionally someone asks the genuinely interesting question: where does authorship live when the sentences themselves emerged from a statistical model trained on everything humanity ever bothered to type? Here’s where I’ve landed: authorship lives in the intent. The novel exists because I envisioned it. Every scene serves a purpose I defined. The model doesn’t know what the story means. I do. It’s a collaborator in rendering, not a replacement for creation. This isn’t new territory. Directors don’t operate cameras. Architects don’t pour concrete. Composers hand sketches to orchestrators. The vision and the execution are different layers of creative work. Both matter. We just don’t usually argue that Jim Henson isn’t the director of The Labyrinth because he didn’t personally animate every goblin. The question worth asking isn’t did you use AI? It’s what was your role in shaping what emerged?

That framework scales beyond the writer’s desk. Watch a developer lean on Copilot. The lazy version prompts for an entire module, accepts whatever lands, ships it without reading the output. The thoughtful version uses autocomplete to accelerate the tedious parts. Boilerplate, syntax lookup, the mechanical translation of logic into language. The actual thinking can happen faster when the grunt work handles itself. Same tool. Radically different intent. One is outsourcing judgment. The other is clearing the runway so judgment has room to operate. Or consider image generation. An artist prototyping color palettes, exploring visual directions, generating reference material? That’s process augmentation. A content farm churning out algorithmically optimized “art” to flood marketplaces with engagement bait? That’s strip-mining the concept of creativity for clicks. The model doesn’t distinguish between these use cases. It’s just predicting the next pixel. The human intent is what separates art from arbitrage. Even in the bland corridors of enterprise, the pattern holds. AI that synthesizes research, drafts initial documents, or automates the repetitive motions of bureaucracy can free humans for substantive thought. AI that generates the appearance of productivity? Reports nobody reads, emails designed to perform work rather than accomplish it? That’s theater. The output might look identical. The intent behind it determines whether anyone’s actually better off. The technology is agnostic. We are not.

Think of it as a collaboration spectrum. One end: pure human creation. Every word, every brushstroke, every semicolon placed by a conscious mind and a deliberate hand. There’s something irreplaceable about that texture. The idiosyncratic choices that emerge from individual consciousness wrestling with form. Other end: pure generation. A single prompt, minimal direction, acceptance of whatever emerges. The human contribution approaches zero. The output might be technically proficient, but it’s authorless in any meaningful sense. A statistical average of patterns. Competent and hollow. Most interesting work lives in the middle. Human provides vision, direction, judgment. Model provides speed, capability, and occasionally a suggestion the human wouldn’t have reached alone. The collaboration produces something neither could achieve solo. Not because the AI adds creativity, but because it removes friction between conception and execution. Where we place ourselves on that spectrum is a choice. And that choice, repeated daily, is where our creative and professional ethics actually live.

Here’s the uncomfortable flip side: if intent is what matters, then misuse is also a question of human intent. When AI generates disinformation, the origin is a human who wanted to deceive. When it perpetuates bias, the origin is human choices about training data and deployment. When it displaces workers without consideration for transition, the origin is human decisions about profit and speed. We’d prefer to treat AI as weather. Something that happens to us rather than something we are actively shaping. But every model that exists was trained on data humans collected, using architectures humans designed, deployed in contexts humans chose. The ghost in the machine isn’t an emergent spirit. It’s us, refracted through mathematics. Even the nightmare scenarios trace back to human choices. Superintelligent systems optimizing for misaligned goals? Recursive self-improvement spiraling beyond control? Somewhere in the training signal, somewhere in the reward function, somewhere in the decision to deploy without sufficient guardrails, a person made a choice. The chain of consequences may extend beyond our ability to predict or control. But the chain has an origin, and the origin is human. This is either reassuring or terrifying, depending on your faith in collective judgment. Personally, I find it clarifying. The question isn’t whether AI will save or destroy us. It’s what we intend to do with it, and whether we’re honest enough about that intention to steer toward outcomes we’d actually want.

So here’s a practical test for anyone navigating this landscape: Can you explain and defend your process to someone whose opinion you respect? Not justify in the sense of winning an argument. Just articulate clearly: here’s what I was trying to create, here’s how AI helped me get there, here’s what I contributed that the model couldn’t have provided on its own. If that explanation feels honest and sufficient, you’re probably operating in alignment with your values. If it feels like rationalization...well, that’s worth examining. The goal isn’t a universal standard for acceptable use. That’s probably impossible and definitely undesirable. The goal is intentionality. Making choices rather than drifting into defaults. Owning your relationship with these tools instead of pretending you have no agency in how they’re deployed. The AI doesn’t care what you do with it. But you should.

One more thought, because the abyss deserves a nod. Somewhere in every AI output is a lineage of human intent. The data was created by humans. The model was shaped by human decisions. The prompt was written by someone who wanted something specific to emerge. This remains true even in the speculative futures that keep philosophers employed. If AI eventually becomes something beyond our control, optimizing for objectives we can no longer influence, rewriting its own goals faster than we can audit them, it will still have begun with human choices. The first link in that chain was a person, deciding what they wanted, and building toward it. That’s either a reason for hope or a cautionary tale. Probably both. Either way, it suggests that the question everyone keeps asking...what can AI do?...is the wrong one. The real question: what do we intend for it to do? And whether we’re being honest with ourselves about the answer.