Free users are costly for OpenAI. One of its biggest challenges throughout this change will be introducing ads at scale, without deteriorating trust or pushing users to competing chatbots, like Google’s Gemini or Anthropic’s Claude.

“It’s not going to be easy for ChatGPT to, let’s say, erode the quality of the experience without losing a lot of users,” says Stefano Puntoni, a marketing professor at Wharton who researches generative AI. Neither of OpenAI’s main competitors features sponsored ad buttons in the outputs, though Google recently said it’s not ruling it out.

Sometimes, when I asked a question with a name brand in the prompt, like DoorDash or Netflix, the ad below the answer was for one of the company’s direct competitors. Toubia describes this as “poaching” and says this technique is a longtime staple of digital advertising in search engines. “That definitely has been a key engine behind the growth of online advertising,” he says. “It seems like it’s going to be the case also with [large language model] advertising.”

Right now, OpenAI is hiring for multiple positions, from software engineers to marketing leaders, to work on this core integration of ads. One of the open positions on OpenAI’s site is for a “product marketing lead, advertising,” and part of the role’s responsibilities is to “identify product risk areas (e.g., performance, safety, policy, trust) and drive cross-functional plans to mitigate them.” Ads come with risks, and how they’re executed will shape the future of the company.

If I exclusively used the free tier of ChatGPT, the introduction of these ads would have me considering other AI tools. Even with OpenAI’s explicit ad policies, there’s an aura of surveillance these ads introduce to the user experience, which is more personal with a chatbot compared to the traditional Google Search experience. While I know advertisers can’t currently influence ChatGPT’s outputs or see my chats, the incessant ads below answers made the conversations seem less private, and I felt hyperaware of the personal data I was sharing with this bot.

After this limited rollout in the US, OpenAI will move to the next phase. “We’re seeing no impact on consumer trust metrics, low dismissal rates of ads, and ongoing improvements in the relevance of ads as we learn from feedback. These positive signals support moving into the next phase of our pilot,” reads an update on OpenAI’s website dated March 26. The company is expected to expand this ad push to Canada, Australia, and New Zealand.

Everything is on the line for OpenAI as this ad rollout spreads to more ChatGPT users.

“The worst thing the company could do is to go in very aggressively, to do it in a way that’s basically maximizing conversions and referrals but at the same time undermining people’s confidence and trust in those recommendations,” Puntoni says. “Then, that will basically be the end of it, because there’s no point using a chatbot you don’t trust.”