OpenAI’s Sam Altman used to hate ads. Now he’s selling them.

• ChatGPT has hundreds of millions of users, and only a small percentage of them pay to use the service.
• But OpenAI CEO Sam Altman wants to turn those free users into revenue-generating users by showing them ads.
• This move has been a long time coming, and has real risk: What if users think ChatGPT’s answers are influenced by the ads it sells?

OpenAI’s Sam Altman used to think ads were gross.

Now he’s rolling them out.

OpenAI has formally announced that it will start showing ads to some US users, starting Monday. The ads will show up for some of ChatGPT’s free users, as well as some users who’ve subscribed to ChatGPT Go, a new $8-a-month tier the company rolled out last month.

Here’s a mockup:

OpenAI says it will serve ads to you based on the conversation you’re currently having with ChatGPT, as well as previous queries and chats. It will also factor in whether you’ve engaged with — or hidden —other ads it has shown you.

For now, the company says, it won’t use data about what you do outside of ChatGPT to target ads inside the service. But I’d be surprised if they don’t do that eventually, since just about every other big internet ad platform uses those signals.

The fact that ChatGPT is launching ads isn’t news: The company has been circling the idea for months, and last month it formally announced that ads would be coming to the service.

And that news became the subject of a back-and-forth between OpenAI and rival Anthropic, which included a testy social media post from Altman and a cheeky Super Bowl ad from Anthropic underlining the possible trust issues OpenAI may have once ads intermingle with “organic” results on the service.

Now we’ll see how these things actually look, and work — and what users and advertisers think of them.