A while back I read a papal encyclical about artificial intelligence. Forty-two thousand words. Buried in there was the best argument for everything I do (and also the sharpest case against it).
The warning it contained has a name. Babel Syndrome. The idea that a machine can take the whole mystery of who you are and flatten it into data and performance. Effectively reducing a person to a row in a database.
My stomach dropped a little. Okay, a lot.
Because isn't that exactly what I help people do? I spend my working life helping business owners make themselves legible to machines. Structure your expertise so the AI engines can read it, understand what you're the expert in, and recommend you when somebody goes looking. Turn your authority into something a machine can find.
Which sounds like Babel, except with a business model attached to it.
Then it flipped for me.
The machine is going to turn you into data either way.
It's already scraping. Already summarizing. Already deciding who gets recommended and who gets forgotten when somebody types a question at eleven at night, and it's doing all of that whether or not you participate, whether or not you consented, whether or not you have any idea it's happening.
You don't get to opt out of being data. You only get to decide who holds the pen.
Babel is what happens when you do nothing, and the algorithm writes your entry on its own terms. Generic. Interchangeable. A commodity it can swap out for the next person who does roughly what you do.
The other option in the encyclical is Nehemiah, which is building on purpose, deliberately, choosing what gets built and how it gets built.
I felt pretty good about that for about… a day.
Then I noticed I'd only answered the smallest version of the question.
The Pope's objection is about human dignity, and it's a serious one, but that isn't the objection I hear from people in my hometown or from people I engage with online.
When I'm cornered about my role or about AI, no one is telling me I'm flattening the human person into a database row.
They say the data centers are drinking the water. They say their power bill went up. They say their daughter finished law school and there's nothing left for her to do.
"Who holds the pen" is a lovely sentence and it doesn't obviously answer a single one of those.
So I went and worked out the rest of it, and it took me longer than I expected it to, because it turned out there were five things I had to settle before I could stand behind my own work with a straight face.
- Whether I'm even entitled to an opinion.
- Which of the frightening numbers hold up.
- What this fight is about underneath.
- What I would ask for if anybody asked me.
- And what I'd say to somebody whose kid did everything right and got shut out anyway.
I'll take them in the order they came to me, because the last one turned into something I wasn't expecting, and it's the thing I now believe most strongly out of anything here.
Whether I'm entitled to an opinion
I'll start with the part I'd rather say myself, before somebody else says it about me.
I can sit down with a business owner and build in an afternoon what used to take a team, weeks (if not months), and a budget.
The only reason any of that is possible is that a handful of enormous companies spent hundreds of billions of dollars, plus a great deal of other people's power and water, to develop AI.
I didn't pay for any of that development, and I didn't carry any of the risk…
and I use it every day and charge people money for it.
So I'm not going to tell you my own usage is irrelevant. It isn't. My business exists downstream of the entire buildout and it wouldn't exist without it. Yet it's also beside the point. How much of it I use was never the question.
For a while I thought my AI usage disqualified me from having an opinion, or that it obligated me to defend the whole arrangement as the price of admission.
It does neither one.
Benefiting from something doesn't make me its spokesperson. I can use these tools and name the harm on the same day, and I fully intend to keep doing both.
And once I stopped auditioning for neutrality, the real question got a lot clearer.
It was never whether I benefit. Everybody reading this benefits from things that were built at costs they never consented to and couldn't itemize if they tried.
The electrical grid. Container shipping. The internet itself, for that matter.
So the question isn't whether I benefit, it's what I owe because I do.
I think I owe four things.
The first is to say it out loud instead of laundering it.
The second is to push this capability down to the people the buildout would otherwise skip entirely, which is most of the people I know.
The third is to hold some kind of standard about what it gets spent on.
And the fourth is to back rules with teeth, even the ones that would raise my own costs.
That last one is the test I'd want applied to anybody in this argument, and I include myself in that.
If you want to know where a person really stands, watch what they're willing to ask for. Not what they say.
Which is a fine principle right up until you go to use it, because to ask for anything specific you have to know what's true first. That part was harder than I expected.
Which of the frightening numbers hold up
(The figures in this section are current as of August 2026. Everything else in this piece is the argument, and the argument should age better than the numbers do.)
The water is the part I take most seriously. Roughly two-thirds of the data centers built since 2022 went up in places that were already under water stress.
Nobody stumbled into that.
Those are siting decisions, made deliberately, by people who were looking at a map that showed them exactly where the water was. By people who benefit at the expense of others.
Electricity turned out to be a lot messier than either side wants it to be.
Residential rates were already climbing before the rise of AI, driven mostly by an aging grid that costs a fortune to replace and by weather that keeps getting worse.
Several of the states with the fastest demand growth actually saw their residential rates fall relative to everybody else's. One regional grid, the mid-Atlantic one, is where the studies find data centers genuinely moving prices.
And going forward, as the spare capacity gets used up, the pressure gets real and that whole story changes.
The jobs promise is the weakest claim in the entire pitch.
The whole American data center industry employed something like twenty-three thousand people as of late 2024, and researchers looking for evidence that these facilities grow local tech employment haven't found much of it.
Meanwhile Virginia and Texas each give up somewhere in the neighborhood of a billion dollars a year in tax exemptions to attract data centers.
I know all of that, and I still use AI every day, and I would rather be the person in the room who knows which parts are true than the person those facts eventually get used against.
But the thing that hit me hardest wasn't any single number. It's that almost nowhere in this country is a facility required to tell you how much power or water it uses.
Which is why this argument never resolves. Everybody is standing there shouting estimates at each other and nobody can be proven wrong.
That is starting to change, and it changed while I was writing this. On August 27th the governor of New Jersey signed a law requiring data centers to report their energy use, their peak daily water use, where that water comes from, and what they're running for backup power. Every six months, not once. Environmental groups are already unhappy about a provision limiting public access to some of it, which tells you the fight there has moved on to the details instead of the principle. A version of the same idea is sitting in the U.S. Senate right now.
So it's possible. It just isn't the rule yet, almost anywhere else.
Once I saw that, I started listening to these fights differently. And the anger, it turns out, wasn't pointed where I assumed it was pointed.
What this fight is about underneath
Here's the move to watch for, and once you've seen it you'll see it in every version of this argument.
Whether a data center comes to your community is a question about the terms of a deal.
What the property owner makes. What the county gets out of it. Who pays for the new transmission. What happens to the water table. Who was in the room, and what exactly they signed. That's a governance question, and governance questions have governance answers.
Whether you use AI to run your business is a completely different question with a completely different answer.
Every pitch for the buildout depends on those two staying fused together.
Accept the buildings, and the prosperity follows. But hosting the infrastructure and having the capability are not the same thing, and a county that says yes to a facility doesn't thereby become a place where people can do more with their work.
It becomes a place with a building in it.
Which is exactly why the local people should benefit from a facility whether or not a single one of them ever touches AI. That's not a favor to anybody. It's the deal being an actual deal.
And when I stopped listening for the word "AI" and started listening to what people were angry about, most of it wasn't about the technology at all.
Shell companies buying up land under other names. Nondisclosure agreements with county officials. Tax abatements nobody got to vote on. Residents finding out what got approved after it was already approved.
That's a bad deal, and it's a bad deal about land and water and tax policy.
I would join that fight tomorrow, and joining it costs me nothing at all, because it was never an argument about whether this technology should exist.
There absolutely is a way to do this responsibly, ethically, and in an environmentally sound way. It has to be thought out. It has to be planned. It needs real experts weighing in and real community insight taken into account, before the signing rather than after it.
Which is the point where "it should be monitored" stopped being good enough for me. Monitored by whom? And what happens when the monitoring turns something up?
What I'd ask for, and I'd pay for it
Three things. I can say all three of them in one breath, and that's the difference between having a position and just having a feeling about something.
Disclosure with teeth. Every facility reports its power and water draw publicly, on a schedule, the way any other large industrial user has to. What we have almost everywhere is voluntary pledges, and voluntary pledges are unenforceable by design.
Ratepayers don't fund the buildout. If a company needs new transmission built for its facility, that company pays for the transmission, instead of the grandmother three counties over finding it on her bill.
Water before permits. No new siting in a water-stressed region without a water plan the community gets to see before the deal is signed.
Every one of those makes my own inputs more expensive.
I want them anyway, and I'd rather be judged on that than on anything else in this essay.
What none of it does is help anybody's child. Buildings and water tables are the easy half of this, because at least there's a permit and a hearing and somebody's signature you can point at.
The other half doesn't come with a hearing you can show up to.
The kid who did everything right
When somebody tells me their kid graduated into a market where the first rung got sawed off, I don't argue with them. I agree with them. It's concerning, and I don't have the whole answer.
I understand the disdain, especially coming from younger people.
They were told their entire lives to go to college, take on the debt, become a lawyer, and now the entry-level clerk work that new lawyers cut their teeth on is getting consumed by AI.
Pretending that isn't happening would be insulting to anybody living through it.
What I won't do is hand that problem back to the twenty-six-year-old with the loans and call it resilience.
She didn't cause any of this, and asking the youngest person in the building to absorb it is the same move I object to when it comes from people with a lot more power than me.
Besides which, she isn't the one who gets to decide. The firms decide. And the firms are who I advise.
So this is what I tell them. Hire the law grad. Task her with implementing AI into the practice. Put her in charge of ideating what the company is going to look like and how it will operate itself in ten years, and give her a mandate to figure that out in a way that doesn't reduce the workforce but expands it right alongside the innovation.
Clerking was never anybody's dream job. It's the old ladder and how things have always been done.
What the law grad wants is experience and a way up, and losing the old way of doing things isn't the same problem as losing the point of the job.
There's a version of this where she gets both, and gets them faster, doing work that matters.
Whether she gets that version is up to the law firm.
Which means every company is an innovation company now.
Not as a slogan about disruption, but as a plain statement about who's on your payroll and what you've actually asked them to do.
And it was somewhere around there, with all five of these sitting on the table at the same time, that I finally saw the thing I'd been missing.
What I noticed with all of it laid out
Every one of those five is the same shape.
A county that doesn't engage with a data center proposal still ends up with a data center decision. It just gets one that somebody else authored, in a room the county wasn't invited to.
A profession that doesn't decide what happens to its entry-level roles still gets an answer, and the answer gets written by whoever moved fastest.
A community that decides this isn't its world still lives inside an economy shaped by it, with none of the upside and all of the electricity bill.
And a business owner who never authors her own authority still has an entry in the machine. She just didn't write it.
Either you write it, or you get handed whatever the default was.
That's the same choice the encyclical was pointing at, all the way up.
Which reframed the thing I do for a living, and I'll admit it took me a while to say this part out loud.
I've spent years treating AI adoption as a growth question for the businesses I work with, and it is one.
It's also something considerably larger than that.
Widespread small business adoption is a fundamental part of this technology being handled properly.
The chain isn't complicated. It starts with one person. An owner who's thriving with the help of AI talks about it at a chamber lunch, at a dinner table, standing on the sideline at their kid's game.
That creates demand instead of avoidance, which is a completely different atmosphere from the one most communities are sitting in right now.
Then those owners want a say in it.
People who are directly affected by how something gets managed don't much enjoy being railroaded by corporate interests or by government ones, and they'd be showing up with a demand that points in two directions at the same time.
We need access to this, and we need it not to poison us.
That combination is what nobody currently represents.
Government isn't going to hold that voice for us. Big tech holding it isn't going to work either, for reasons I'd hope are obvious.
But if enough business owners are saying I need this, and I need it to be safe, then elected officials have to start paying attention to a groundswell, because a groundswell is about the only thing they reliably pay attention to.
I know the objection here, because it's the one I'd raise. Doesn't adopting it just make you dependent? And isn't a dependent customer a lot easier to ignore than a critic?
Was the automobile a hostage situation? Did the internet take us hostage?
People can opt out of cars and of being online, and most of them don't, because life is better with them.
And look at how the safety arrived. Cars killed people for decades before seat belts were standard, and longer still before the law required anybody to wear one.
You don't get seat belt laws in a country with five hundred cars. You get them once everybody drives, and the harm becomes everybody's problem, and the drivers themselves start demanding them.
We're in the Wild West right now. Almost no restriction, almost no regulation, and that is going to produce some unfortunate consequences in the short term, which will probably force some quick action after the fact.
That's usually how it goes. And it isn't a reason to sit this one out, it's the reason not to.
Because fighting this doesn't stop it. Fighting it is counterproductive to progress in the right direction, which is a different claim than saying resistance is futile.
Opposition forfeits the seat, and the seat is the only place any of this gets decided.
What I tell my family
My daughter is seven. We live rural, the high school option out here isn't that good, and my family is genuinely afraid about what happens when she gets older. They want me to name a plan. A school, a move, some decision they can hold onto.
My honest answer is that I have no idea what the world is going to look like by the time she gets to high school.
If education looks the same in eight years as it does right now, I would be very surprised.
The world is changing, and that includes education and how people get compensated for what they know.
Maybe the novelist stays a novelist, and what changes is that books stop being how the novelist gets paid. I don't have exact answers, and I'm not going to pretend to be a futurist, because I'm not one.
What I can do is raise somebody who authors her way through whatever it turns out to be, instead of somebody waiting around to be placed.
Which is the same answer I give a business owner about their five-year plan, and the same answer I'd give that law graduate, and the same answer I ended up giving myself the day I read forty-two thousand words about Babel and had to sit with whether the work I do was part of the problem.
We are living through an enormous amount of change, and the choice on the table is whether you're willing to get on the train to the new world, or whether you'd rather dig your heels into the way things have always been done.
I'd rather write our future than have it happen to us.
You're already data. The only question left is who's holding the pen.
Where the numbers come from
All figures as of August 2026. If you're reading this much later, check them.
- Data centers in water-stressed places. Close to two-thirds of facilities built or in development since 2022, from Bloomberg's 2025 analysis. Five states account for 72% of those sited in areas of acute water stress. Summary here.
- Electricity rates. That rates were climbing before the AI buildout, that an aging grid and weather are the larger drivers, that several high-growth states saw rates fall, and that the mid-Atlantic PJM grid is where studies find data centers moving prices: E&E News, citing an Edison Electric Institute study.
- Jobs and tax exemptions. Roughly 23,000 people employed in the US data center industry as of late 2024, no clear evidence of local tech employment growth, and about a billion dollars a year each in Virginia and Texas exemptions: Consumer Reports.
- The New Jersey reporting law, signed August 27, 2026: New Jersey Monitor. The federal version is S.4213, the Data Center Water and Energy Transparency Act of 2026.
