ChatGPT is making room for ads from regulated verticals—but marketers aren’t ready to jump in at scale

OpenAI is moving toward accepting more regulated verticals to its ChatGPT ads pilot, but advertisers and their agencies are looking before they leap.

The company, which has been steadily building out its ads offering, began opening its ads pilot more widely to financial services brands in May and has made some recent changes to its ads policies that reflect a loosening of restrictions on highly regulated verticals. But as OpenAI is moving toward expanding the brands that can run ads on ChatGPT, those often-risk-adverse advertisers aren’t necessarily ready to dive in at scale.

Finance clients in particular are “curious with a big dose of caution,” Amy Winger, global client lead at VML, told Marketing Brew. But healthcare brands are looking for ways to get in on the action, she added.

Not so fast

What changes has OpenAI made to its ads policies? In April, the company added language stating that “medical, legal, and financial advice contexts are no longer categorically blocked from ads by default.”. It also added language stating that the company “may approve ads from approved advertisers within the financial services, healthcare and medicine, and legal services categories…manually on a case-by-case basis,” according to a review conducted by Marketing Brew comparing the current ad policies web page with an April 29 iteration of it saved on the Wayback Machine.

The company confirmed that it has updated its ads placement policy to be more precise regarding ads in regulated categories, Taya Christianson, an OpenAI spokesperson, told Marketing Brew. She reiterated that the company isn’t widely including ads in financial, legal, or medical conversations, but noted that ChatGPT may show ads in limited circumstances in those verticals when they make contextual sense. Contexts the company deems non-sensitive for ads in those verticals, she added, may include general queries around exercise or diet, or recommendations for budgeting or financial education tools in response to broad financial questions.

While the adjustments could be intriguing for financial services brands, some of which may already run ads on traditional social platforms like Instagram, not all of them are planning on diving into ChatGPT ads just yet, Winger said, citing that the brands VML represents aren’t making moves at the moment. Already, consumers often heavily consider finance-related purchases, and the industry has strong regulations governing advertising. With that said, VML is keeping an eye on the performance of digital financial players like Robinhood, which has run ads on ChatGPT, to see if ChatGPT ads might make more sense for niche digital financial services companies, she told us.

As consumers “think about financial services in a purely digital way,” some financial services brands may be worried about the possibility that their services could become almost gamified through the platform, or otherwise automatically recommended, Winger said.

“In terms of the interaction between ChatGPT and the potential brand, [there are] possibilities of making decisions that might not be fully vetted when it comes to making the right decision for wealth management,” Winger said. “These are pretty considered purchases, and so making them a little bit more automatic is a little bit worrisome.”

Healthy skepticism

In contrast, healthcare brands seem to be more interested in ChatGPT ads. While pharmaceutical brands aren’t permitted in the pilot yet, Havas has been “pushing” OpenAI to allow them, Holly Dunn, the agency’s managing partner and head of investment and activation, said. She thinks it’s only a matter of time: The company’s move to allow regulated verticals such as financial services is an encouraging step, she said, and OpenAI seems “open to have the conversations” about additional advertising opportunities, she said.

OpenAI has also made some other steps in the healthcare and wellness space, including rolling out a wellness-focused experience inside of ChatGPT, called ChatGPT for Health, in January, as well as releasing a version of ChatGPT called ChatGPT for Clinicians in April.

“We’re [at] a pivotal point where patients are actually becoming practitioners, with the ability to source their own indications and ailments with all this data,” Ray Romero, Havas’s managing partner of client experience, said.

While they wait on engaging with consumers through ChatGPT ads, many pharmaceutical brands are already advertising on physician-specific medical GPTs, AI offerings where healthcare professionals usually have to apply to gain access, and where information is trained on specific professional data, Dunn said. Havas healthcare clients are running ads on OpenEvidence, Doximity, and DougallGPT, according to the agency.

While there’s pharmaceutical brand interest in ChatGPT ads, there are certain issues about platform trust that those in the space may like to see be addressed, too, Romero said.

“ChatGPT has a little bit of a hill to climb; they started in a world where they’re sourcing external third-party data [and] the outputs of whatever searches you get on there were either super strong or sometimes questioned,” he said. “You get into a world where there’s a lot of concerns with AI around hallucinations or false claims, and those are the things that you really have to be careful of as you start thinking about AI within the healthcare space.”

Two steps forward, one step back?

As there are some more openings for advertisers on OpenAI, the platform is also strengthening its restrictions around ads in certain verticals and conversations, per our Wayback Machine review. The company added “mental and personal health conversations” as an additional sensitive context against which ads can’t run; a previous iteration of the page did not list “personal health conversations” as sensitive.

OpenAI also added a “review scope” section to its policy that details how ads are evaluated, which involves examining ad creative and landing pages to ensure “compliance with our policies.”

Christianson reiterated that sensitive conversations and related contexts are still barred from ads.