These are the biggest concerns holding back marketers’ adoption of AI

But as they rush to embrace the possibilities of generative AI and agentic workflows, they are frequently encountering a complex web of legal, ethical and practical concerns.

The World Federation of Advertisers (WFA) has responded by creating what has become a substantial AI governance community to help marketing leaders navigate these waters. Gabrielle Robitaille, director of policy at the WFA, leads this initiative, which has quickly grown to more than 700 people representing 135 of the WFA’s 148 member brands. The group includes everyone from CMOs and chief brand officers to global heads of media, legal counsels and chief data privacy officers.

“It was probably the first time since GDPR that brands were like, ‘OK, we have to speak to legal, we have to speak to procurement, we have to speak to policy. We all need to be on the same page about this stuff. Otherwise, we’re going to get it wrong and we’re going to miss out on the huge growth opportunities,’” Robitaille says.

Legal roadblocks are the top challenge

A steering team encompassing marketers from companies such as Mars, L’Oréal, AXA and Teva Pharmaceuticals helps to ensure the group is tackling the most pressing issues. At their Chatham House rules in-person workshops, such as at the recent Global Marketer Week in Brussels, one subject has consistently been on the agenda – legal.

“Legal roadblocks are the number one kind of challenge to scaling AI right now,” says Robitaille. “If you dig into those legal challenges, it’s IP – how to protect your brand IP – it’s data protection and privacy.”

The WFA has published IP guidance to help brands understand and mitigate their risks when using AI in marketing contexts. But using AI appropriately, especially in sensitive sectors that are highly regulated, is a matter of team culture as much as corporate governance.

Take, for example, the common concern that Robitaille would hear about teams inputting confidential company information into large language models such as Chat GPT. “They were like, ‘We don’t know what to do. We don’t know how to equip teams with the knowledge that they can’t share this information.’”

For most brands concerned about this scenario, the typical response has been to use AI only within a walled garden of enterprise-level tools deemed safe by legal and IT departments. But that, of course, slows down progress. “We were finding out that all of these efficiencies that AI promised actually weren’t happening because there were so many checks to go through before you could start using [it],” Robitaille says.

It also stymies marketers’ personal development when it comes to familiarizing themselves with the technology and its capabilities. As Robitaille says: “There are new tools and models coming all the time and marketers just want to try all of them, right?”

Some brands have developed innovative solutions, like automated intranets that streamline the approval process for new AI tools, directing users to pre-approved alternatives when available and only requiring full security reviews when absolutely necessary.

But while that might help them get their own house in order, it doesn’t assuage anxieties about how their agency partners are using the technology – and that’s another big concern that the WFA’s AI community has been addressing.

Gabrielle Robitaille speaking during the WFA’s Global Marketer Week

Agency anxieties need to be overcome

Advertising agencies have been aggressively building their AI offerings to convince clients that they can take all of this stress away from them and allow them to merely enjoy the ‘faster, better, cheaper’ benefits the tools will supposedly bring.

For the most part, this message has been landing, with brands happy for their partners to take on the upfront cost and complexity of finding the most suitable AI platforms rather than attempting to do this themselves.

But there’s a kicker. “Our research found that one of the biggest concerns that brands have after the legal risks was how agencies are using AI and the lack of transparency,” Robitaille says. “We [also] saw quite a few media reports of big brands introducing clauses that were like, ‘You cannot use AI without having consent from us first.’”

The WFA developed AI contract best practice guidance to help brands mitigate risks while still allowing agencies to innovate. Interestingly, rather than looking to cut agency costs through AI, Robitaille notes that brands are, for now at least, “more inclined to pay their agencies more for AI expertise because they really want to invest in making sure they’ve got people that know what they’re doing.”

With cost savings frequently touted as a key part of AI’s appeal, it remains to be seen how long this will last. But “right now,” Robitaille says, “they’re actually willing to invest more in that agency talent than to try and cut costs.”

Ethical concerns are on marketers’ minds

The impact AI is having, and will have, on partners and talent is being deeply considered by the group, according to Robitaille. But there are other ethical issues that they are wrestling with too, such as sustainability.

“I think there was this misconception around a lot of brands that, ‘Hey, we can cut production, we’re going to save on loads of carbon footprint.’ But actually, it might be worse than what we were doing before.”

Bias – both in terms of team inputs and AI creative outputs – is another area where education is required. Robitaille says brands are increasingly experimenting with tools that are coming to market to help them identify bias in outputs, but, ultimately, it will be down to “individual company policies that will decide where they want to be on that risk level.”

The WFA isn’t trying to establish rigid standards around AI use. Instead, it is focused on raising awareness of best practices while recognizing that brands will take different approaches based on their own ethical stances.

“We recognize that not all brands are going to be taking similar positions on these things. Some brands are far more comfortable with using synthetic humans in their creative and others have said we will never go there because we see it as unethical,” explains Robitaille.

The community works by crowdsourcing insights from what different companies are doing, extracting the most valuable ideas and then putting that into practical guidelines that brands can apply.

It’s not just about policy – it’s also about (internal) politics

The question of who should lead AI initiatives within organizations remains contentious.

“I don’t think there’s best practice yet for who owns AI internally within the business,” Robitaille admits. “I think, naturally, it has landed with IT and tech officers who don’t necessarily understand the business side of things and the business value of AI.”

While marketing isn’t traditionally where technology ownership would fall, marketing applications represent some of the most promising AI use cases. This has created internal “frustrations and tensions” that many organizations are still working through.

Diplomacy is key. “What we see [being most likely] is that it is still often owned by IT, but that there are these governance structures in place where marketing will be part of the discussions, or that there will be separate marketing AI board that will be dealing with the issue.”

The biggest wins aren’t always the most glamorous

For companies that have managed to overcome these hurdles, many of the most valuable use cases so far are the ones consumers never see. “What is having a real impact is those internal use cases that people don’t see, which aren’t necessarily the most exciting ones, but it’s like media campaign reporting, for example,” she says. “If you’ve got your data architecture right... you’re able to extract these amazing insights that help you further optimize your media buying and media planning.”

Ultimately, however – and for all the AI hype – many marketers are still finding their feet.

“We ran a poll yesterday during one of our sessions and 66% of the audience... said that their brand was just in the developing stage, just trying to figure out what their AI strategy would be,” Robitaille shares.

As marketers are discovering at their own pace, the promise of AI is real – but so too are the growing pains.