Okay so I was going to write about something else today and then this OpenAI thing crossed my feed this morning and I had to change plans, because I've apparently been waiting fourteen years to say "I told you so" to a chatbot.
Here's the news part: OpenAI announced yesterday that it's going to start testing visual ads next to ChatGPT's image generation results. Not inside the image, they're careful to say that, the ad sits beside the picture you asked for, labeled as an ad, and supposedly doesn't touch the actual answer ChatGPT gives you. It's a limited test, US-only for now, rolling out to a small group of advertisers later this month. The ads are meant to show "product inspiration," how something's used, that kind of soft-sell stuff, not some flashing banner for car insurance. OpenAI's also lined up Haus, Measured, and WorkMagic to run geo-based measurement experiments for advertisers, and they're building brand-safety pilots with DoubleVerify and Integral Ad Science, which tells you this isn't a quick hack somebody shipped on a Friday. There's a whole ad-ops machine spinning up behind it.
This isn't actually new for OpenAI, to be fair. They put ads into the free and low-cost ChatGPT tiers earlier this year and pushed that into India back in August. The image generation thing is just the next surface. And once it's rolled out more broadly, the plan is obviously to point that machine at all 1.2 billion weekly ChatGPT users, because that number is the whole reason any of this is happening.
I don't think this is some huge scandal, and I'm not going to pretend I'm shocked. Of course it's ads. Running a service for over a billion weekly users costs an absurd amount of money and subscriptions alone were never going to cover it, not with how much compute image generation burns through. I get the math. But I still have a very specific, very personal reaction to this, which is: I've seen this exact movie before, and I was the projectionist.
Back in like 2012, maybe early 2013, I bolted AdSense onto this very blog. Little rectangle up top, one in the sidebar, very tasteful, very "I am a serious tech blogger now." I think my first payout check was something like four dollars and change, and it took months to get there because Google's threshold back then felt designed to mock you. I kept the ads up for maybe a year and a half before I ripped them out, partly because they were ugly next to my (also ugly, lets be honest) layout, and partly because watching Google's ad network decide what belonged next to my writing made me a little nauseous. One week it was fine, tech stuff, dev tools. The next week it was weight loss pills. There's no "label" that makes that feel better, it just feels like someone else is standing in your living room talking to your guests.
So when I read "the ads will be clearly labeled and won't influence the answers ChatGPT provides," sure, I believe that's technically true today, on October 5th, 2026, with this specific test. What I don't believe is that it stays this contained forever. It never does. Free tier gets ads. Then it's a region. Then it's a feature. Then it's everywhere, and the labeling gets smaller, and the line between "ad" and "suggestion baked into the answer" gets fuzzier because that line is worth money to blur. I'm not saying OpenAI is lying to us right now, I'm saying incentives don't care what anyone promised in October.
Picture it: you ask ChatGPT for a picture of your dog in a cozy reading nook, because you want a header image for something, and right there next to your dog is a Chewy ad for the exact beanbag you thought you dreamed up yourself. Mildly funny the first time. Deeply unsettling the fortieth time, when you can't tell anymore whether you wanted that beanbag or whether something nudged you toward wanting it.
Anyway. I don't have a tidy moral here, and I'm genuinely curious whether the brand-safety pilots with DV and IAS end up mattering or whether that's just vendor theater. I'll probably still use ChatGPT for images next week regardless, because it's useful and I'm not a purist, I just run a blog that's had actual honest-to-god banner ads on it, so I know exactly how these stories tend to go from here. Keep an eye on it. I will.