Everyone is treating this as the moment ChatGPT started its slow slide into enshittification. I think that reading is backwards. OpenAI is bringing ads to ChatGPT’s Free and Go tiers in India, and if you care about AI assistants staying available to people who cannot drop twenty dollars a month on a chatbot, this is closer to good news than bad.
Let me explain what actually happened, and then why my take is what it is.
What we know
The reporting across TechCrunch, Business Standard, Seeking Alpha, and others lines up on the basics. OpenAI is starting to show ads to ChatGPT users on the Free and Go tiers in India. Go is the cheaper paid tier, so this is not only a free-user thing — some paying users will see ads too. Coverage from Pluang points to roughly 50 brands in the initial rollout, served through a new ad manager built for the purpose. Inc. framed the scale bluntly: around 100 million users will start seeing ads in ChatGPT. And OpenAI’s line is that the answers themselves won’t change.
Separately, the company is expanding operations in Brazil. India first, Brazil next. That sequencing tells you something.
Why India, and why this is not an insult
A lot of the commentary I’ve seen treats “ads launch in India first” as OpenAI testing its most annoying idea on users it respects least. I’d flip it. India is one of ChatGPT’s largest user bases, and it is a market where a twenty-dollar monthly subscription is a genuinely different proposition than it is in San Francisco. If the only sustainable way to run a large model is to charge every user a Western software price, then most of the world gets a worse assistant or no assistant at all.
Ads are how the internet solved that problem for search, maps, email, and video. They are an imperfect solution with real costs. They are also the reason a student in Pune and a product manager in Palo Alto can use the same search engine. Testing an ad model in the market where price sensitivity matters most is not disrespect. It’s where the model has to work if it’s going to work anywhere.
The part I’m actually watching
The claim that matters most is OpenAI’s promise that answers won’t change. That is the whole ballgame, and it’s a harder promise to keep than it sounds.
With a search engine, you can see the seam. The ads sit in a box at the top, labeled, visually separate from the ten blue links below. You know which part someone paid for. An AI assistant does not have that architecture. It gives you one answer in one voice, and that voice is the reason you trust it. If a sponsored preference ever leaks into the reasoning rather than sitting beside it, you would have no way to tell. There’s no seam to inspect.
So the questions I’d want answered, as someone who explains this stuff to non-technical readers for a living:
- Are ads clearly labeled and visually separate from the assistant’s actual response?
- Does anything a brand pays for influence what the model recommends, ranks, or mentions?
- What information about your conversation is used to pick which ad you see?
- Can Go subscribers pay a bit more to turn ads off, or is the ad tier now permanent for them?
Those four answers will tell you more about OpenAI’s direction than any launch post will.
What this means if you just use ChatGPT to get things done
Practically, in the near term, probably not much. If you’re in India on Free or Go, you’ll start seeing sponsored content in the interface. Fifty brands is a small pool, which suggests a careful start rather than a flood.
The habit worth building now is a small one: when an assistant recommends a specific product, brand, or service, treat that as a starting point rather than a verdict. Ask why. Ask for alternatives. Ask what the tradeoffs are. That was already good practice before ads existed, because models have always carried whatever biases sat in their training data. Ads just make the incentive visible enough that people finally act on advice they should have been following anyway.
The honest version
I’m not going to pretend this is purely user-friendly. OpenAI needs revenue, running these models is expensive, and an ad business gives the company a new reason to want your attention rather than just your satisfaction. Those incentives can drift. They have drifted at every company that went this route before.
But the alternative to an ad-supported tier isn’t a pristine ad-free assistant for all. It’s a paywall that a lot of people can’t clear. Given those two options, I’d rather have the ads and stay loud about where the line is: sell me a placement next to the answer, never inside it. India is where that line gets drawn first. Brazil is next. The rest of us should pay attention to how it holds.
🕒 Published: