In October 2024, during a fireside chat at Harvard, OpenAI CEO Sam Altman said that the idea of intermingling ads and AI was “uniquely unsettling” to him.
“I will disclose, just as a personal bias, that I hate ads,” he said. “I think ads were important to give the early internet a business model, but they do sort of somewhat fundamentally misalign a user’s incentives with the with the company providing the service.”
He went on to say that he viewed ads as a “last resort” business model, and played into OpenAI’s nonprofit origins, saying that the company’s plan was to get “rich” people to keep paying for access to its AI products, so people who couldn’t afford to pay could still use things like ChatGPT for free. “You’ll see us do a lot more to make the free tier much better over time,” he promised.
Subscribe to get the latest creator news
Now, just over a year later, OpenAI is putting ads in the zero-cost and low-priced tiers of ChatGPT, while the “rich people” who pay higher subscription costs will remain ad-free.
If Altman truly meant it when he said ads are OpenAI’s “last resort” business model, then this is a troubling turn for the company that’s at the forefront of the AI bubble and is “looking at [capital expenditure] commitments of about $1.4 trillion over the next 8 years,” mostly in data centers, according to Altman. This is also coming after he sent out a “code red” memo telling engineers to improve ChatGPT asap in the wake of Google debuting Veo 3.
OpenAI hasn’t directly addressed the significant discrepancy between Altman’s speech and its new plans. In a company blog post, its incoming CEO of Applications, Fidji Simo, pitched ads in an altruistic light similar to Altman’s comments about rich and poor people.
“AI is reaching a point where everyone can have a personal super-assistant that helps them learn and do almost anything. Who gets access to that level of intelligence will shape whether AI expands opportunity or reinforces the same divides,” she wrote.
She went on to explain that only logged-in free users and users paying $8/month for its “Go” tier will see ads.
Ads will be hyper-targeted…kind of. “To start, we plan to test ads at the bottom of answers in ChatGPT when there’s a relevant sponsored product or service based on your current conversation,” Simo wrote (so users will get ads for something they’re already interested in). “Ads will be clearly labeled and separated from the organic answer. You’ll be able to learn more about why you’re seeing that ad, or dismiss any ad and tell us why.”
The reason only logged-in users will see ads is because OpenAI is trying to head off any child privacy issues. Even if someone is logged in, OpenAI said it won’t serve ads to anyone it suspects is under 18. It’s also not going to run ads against “sensitive or regulated topics like health, mental health or politics,” Simo said. (We imagine an ad in the middle of a mental health conversation would be quite the exhibit if yet another wrongful death lawsuit is filed.)
OpenAI already has plans to take these ads further: “Given what AI can do, we’re excited to develop new experiences over time that people find more helpful and relevant than any other ads,” Simo said, adding that because of ChatGPT’s conversational interface, “soon you might see an ad and be able to directly ask the questions you need to make a purchase decision.”
As for what will be marketed in those ads, we’re not sure–but we do know brands will probably be paying a lot to secure slots. According to The Information, OpenAI is asking marketers to pay $60 per 1,000 views, a CPM that’s triple what competitor Meta charges on average.
Brands will have to temper their expectations on returns, too: OpenAI reportedly will not track user purchases or more invasive data the way YouTube and Facebook do; instead, it’ll give marketers generic performance metrics like ad views and clicks.
That’s likely because OpenAI promised in its announcement that despite selling ads, it’ll “never sell your data to advertisers.” We mentioned above that ads will be sort of hyper-targeted. We say sort of because if OpenAI isn’t selling data to advertisers, there’s only so much targeting it can do.
The high price + lack of data is another ingredient in this contradictory soup. Altman hates ads, but is running them. OpenAI is spending hundreds of billions, but might foresee a financial hole. Advertisers can pay more to get fewer data points.
Will this work out? We’ll see…
