‘Don’t trust those models’: How marketers are thinking about the next era of transparency
At the Advertising Research Foundation’s AudiencexScience conference in March, execs talked collaboration and keeping up with AI tools.
• 4 min read
Don’t understand how the AI models running ads for your brand work? That could be cause for concern, according to Bob Lord, president of Horizon Media Holdings and interim CEO of Horizon Global.
“There are going to be models that are using AI agents and communicating with one another. I would say to a marketer, don’t trust those models,” he said onstage at the Advertising Research Foundation’s (ARF) AudiencexScience conference in New York in March. “Just don’t give Mark Zuckerberg $100 million and let him do what he wants with it.”
It’s not just Lord who’s emphasizing the importance of transparency in the age of AI. As AI continues to transform marketing, increasing transparency into how the tech works will be key, according to several marketers who appeared onstage.
Sharing is caring? Open-source tech, whether that’s in the world of AI or ad tech, can offer a way to promote transparency through collaboration, Lord said onstage. He noted that those who don’t know how to code but have ideas are now able to “vibecode” using tools like Claude. The increased collaboration going on today stands in contrast to an earlier era, when “software companies in particular would go to the open-source world, take code out, and create a business around it, and then create a proprietary model, and they [would] never give code back into that open-source world,” Lord said.
“The problem with that is, you’re only as good as what your company is,” Lord said. “Your proprietary technology, or your proprietary answer, will only last for a period of time because you’re not tapping into the global innovations that are happening.”
But adland still faces certain barriers when it comes to technological innovation. For one, Lord said, agencies tend to create proprietary tech, which he believes can limit innovation by potentially narrowing options for clients. In the measurement world, embracing an open ecosystem may also be key to staying ahead of the curve; sticking with one tool could cause marketers to “lose out in the long run,” he said.An agency’s stack “may be called open, but it sure the hell isn’t open,” Lord said. “You’ve got to think about, ‘Is it interoperable? Can I actually share the data in different ways?’”
Get marketing news you'll actually want to read
Marketing Brew informs marketing pros of the latest on brand strategy, social media, and ad tech via our weekday newsletter, virtual events, marketing conferences, and digital guides.
By subscribing, you accept our Terms & Privacy Policy.
There is some movement toward collaboration, particularly in the realm of clean rooms, which have allowed for increased data sharing. Google, for one, made its open-source marketing mix model, Meridian, widely available in January 2025, which Lord said onstage was hard to imagine happening even five years ago.
But as AI tools take off, their complexity could become another barrier to transparency, Scott McDonald, ARF president and CEO said. The ARF, he added, might begin to look at building frameworks for “understanding when to believe an AI solution versus something that we might want to pressure-test a bit,” he said onstage.
“Even if you had a great spotlight into the algorithm, many of the developers of AI [models] aren’t exactly sure how they work because at some point, the machines are responding to each other and running off on their own,” he said.
Breaking it all down: To that end, “explainability and training” are also key to transparency, Lord said. Data scientists should not only be able to run code on their own to understand how an LLM makes decisions, but also be able to access information on what data the AI has been trained on to evaluate potential biases.
In the measurement world, explainability is also becoming increasingly paramount, Pete Doe, chief research officer at Nielsen, said onstage. Nielsen has been working to help clients better understand the difference between its previous panel-only currency and its Big Data + Panel offering that rolled out in September as it navigates the currency transition, and as its Big Data + Panel is facing some challenges.
Explainability has “become more and more of a word that is in our vocabulary at Nielsen,” he said. “Explaining the differences between the old measurement and the new is has been a focus for some time, and will continue to be.”
About the author
Jasmine Sheena
Jasmine Sheena is a reporter for Marketing Brew writing about adtech, Big Tech, and streaming.
Get marketing news you'll actually want to read
Marketing Brew informs marketing pros of the latest on brand strategy, social media, and ad tech via our weekday newsletter, virtual events, marketing conferences, and digital guides.
By subscribing, you accept our Terms & Privacy Policy.