Nobody Asks How You Made It When It's Good
The accusations are easy to spot. You used AI, I can tell from the em-dashes. They're also a distraction. The real test is whether anyone stands behind it, and whether it was worth your time to read.
Recently I gave a talk to product and engineering leaders at a large company. Before I sent the deck over, my contact, an exec running the org, looked at it and told me the slides were great. Then he added that clearly I'd made them with Claude.
I said yeah. Obviously.
So when I got up to present, the first thing I told the room was that I'd used AI to design the slides. Yes, all of it. You don't have to study the section dividers looking for tells. Now let's get to the talk.
What I want from slides is support. They should carry the story without distracting from it. I'm not a designer. I just want the talk to stand on its own two legs without messy visuals taking those legs out.
AI gets me there. But in this moment it can become a distraction too, if I let people sit there wondering about it. So I take it off the table in the first thirty seconds.
People only care that you used AI when the work feels careless. The actual problem is not disclosure. The problem is whether the thing was worth the audience's time.
There's a lot of this going around right now. You used AI. I can tell. You used those words. You used em-dashes. Your slides have those layouts.
I have thoughts about it because this is my actual work. I help people adopt these tools. I use them constantly, in what I build and in what I deliver. People are paying me, and I never want them to read "made with AI" as "low effort."
Most people objecting to the formatting have deeper concerns. If someone is genuinely hung up on whether your em-dashes are machine-made, one of two things is probably happening. Either they were never the audience, or the substance is thin enough that the formatting is the most interesting thing in the room.
The hotter reactions are usually about something else entirely. Uncertainty about where the work is heading. Being pushed to engage with a technology people didn't choose and don't trust. A reasonable fear that the person sending the work outsourced the thinking but is still billing you for it. The complaint about your slides is the surface.
Underneath the noise there's a real version of this worry, and it's the one I pay attention to. You are not promised that anyone will consume what you make.
You've sat in meetings where nobody did the reading. You've sent careful documents to executives who skimmed the first paragraph and moved on. Getting busy people to actually take in what you made has always been a challenge.
What's changed is that it's now trivial to generate reams of words. If you're not careful, you build bigger obstacles to being read instead of smaller ones. Three pages where two paragraphs would do. A full memo where a short note would have worked.
So the job got harder, not easier.
One CEO client encapsulates this perfectly. He runs a large, complex organization, and nothing about that engagement matters more than being helpful and not wasting his time. Everything I send him is built around that. I still produce the materials with AI. But I put the value up front, keep it short, and lower the cost of digging in.
You don't earn his attention once. You earn the few minutes he gives you every single time, and now you're earning them against a flood of generated noise from everyone else competing for the same minutes.
That's why the framing around a document matters more than it used to. The summary. The note you write when you hand it over. The signals that communicate that I stand behind every word. Even the ones I didn't touch the keyboard for.
That means, at a minimum, I read it, checked the arguments, and edited it until it accurately represented my points. The cost is my credibility.
When I send a long research document, I lead with the part that is unmistakably mine. My read on it. The framing. Where I landed and why. The research underneath might run long if it needs to, and plenty of those data points might have been gathered by digital hands.
The point of view has to be mine. I stand behind it. That is what actually matters.
I did this before AI too. I led with the point, then put the support underneath. Here's my summary. Here's my conclusion. Here's the material if you want to dig in. Now some of that supporting material might have em-dashes.
The thing people are actually guarding against isn't AI. It's the lazy document. The thing someone forwarded without reading it themselves. AI didn't invent that. It just made it cheap to produce at volume, and people can feel the difference when it lands in their inbox.
Slop isn't a question of whether a machine touched the work. It's a question of whether anyone vouches for it. Make something genuinely useful and nobody asks whether Claude designed your slides. Make garbage and people stop opening your documents, em-dashes or not.