Experts offer tips for using AI for drug advertising and promotion
BALTIMORE – Without guidance from the US Food and Drug Administration (FDA), experts said that companies must develop their own best practices for using artificial intelligence (AI) for pharmaceutical advertising and promotional activities, including creating web, social media, broadcast, or print content.Speaking at RAPS Global Regulatory Strategy Conference last week, Jason Cober, director of regulatory affairs, AI, and digital transformation at ProPharma, and Virginia Foley, CEO of Compliance Forward, discussed effective strategies for using AI in drafting and evaluating promotional activities, as well as circumstances in which it is best not to use AI. The speakers also addressed the proliferation of AI-generated summaries on web and email platforms.
Lack of guidance
Cober, who worked at FDA for 17 years, including eight years serving as lead project manager in the Office of Prescription Drug Promotion (OPDP), said that “there isn’t any guideline which the agency tells us how to use artificial intelligence in ad promotions. So, we kind of have to make inferences about what the agency is thinking.”
In January 2025, FDA issued a guidance focused on the regulation of AI in drug development, though Cober noted that the guidance does not address advertising and promotion. “This is a challenge that we have to grapple with,” he said.
Another challenge facing the industry is the uncertainty surrounding the FDA’s use of AI in the review of promotional materials. In June 2025, the FDA announced that it is using an AI tool called Elsa to help expedite product reviews. However, there are concerns about the agency's lack of transparency regarding how Elsa is utilized in the evaluation of new drugs and promotional content.
"We don’t know how they are using Elsa," said Cober. Yet he said that companies should challenge the agency to clarify how Elsa is being employed in these reviews.
“You can ask them in [pre-investigational new drug application] pre-IND meetings and [new drug application] NDA meetings, are you using artificial intelligence to review my submission? How are you using AI to review our submission? Is there anything I can do in my submission to make your review more efficient?” Cober asked.
He added that “they may not answer your questions, but that doesn’t mean your questions do not come up. I can tell you from my experience at OPDP that the questions I got really informed our decision making, it informed our insights as to what industry struggling with … They won’t be put off by it. Trust me. A lot of companies do not want to ask those questions and honestly, the companies that did ask those questions, we were able to be more responsive, and we were able to help them more.”
When AI to use AI, or not
Cober then discussed some best practices, highlighting what AI excels at and what it struggles with in the realm of advertising and promotion.
He mentioned that one of his responsibilities at ProPharma involves monitoring and identifying trends in FDA untitled letters for promotional activities. He indicated that AI has been useful in analyzing trends from 84 recent untitled letters issued by OPDP related to violations in prescription drug promotions.
Cober said that “this is where we can use AI to summarize and to understand the key phrases and to explain what these trends means.”
He added that “if you are preparing a report for your manager and you want to share the trends, this is when AI can be very powerful and useful.”
However, he said there are “pitfalls” associated with using AI. He provided an example of how he utilized two programs, ChatGPT and Claude, to evaluate a warning letter. He noted that these tools struggled to capture the nuances of the letter. For instance, certain text in the letter was underlined for emphasis, but neither program picked up this formatting change.
“We have to know when it is good to use AI and when it does not. It doesn’t understand everything. The devil is in the details,” Cober said.
Foley said that using AI tools has been helpful in shaping copy for promotional advertising and helps her get ahead of the huge volume of copy that must be disseminated across multiple media channels.
“We are drowning in ads and promotional advertising. I’m sure you are as well. There are more channels out there, TikTok videos and emails, and localized and personalized ads going to specific providers, and the amount of volume that is increasing across channels is daunting, and we are all familiar with regulatory budgets, and they are not increasing to handle this capacity,” she said.
Foley explained that she uses a type of Generative AI tool known as atomic content to streamline her work and create brand-compliant content. Atomic content consists of self-contained messaging units that can be assembled into compliant content across various channels, including social media, television advertising, and emails. This approach breaks information into small, reusable components that can be leveraged across different media platforms.
RAPS Global Regulatory Strategy Conference
