I’m writing today’s newsletter from the Cannes Lions Festival of Creativity, the annual confab where advertising and marketing honchos meet on the Riviera to sip rosé, hand out awards, and talk about the future. This year, AI was top of mind for many attendees, as it has been for executives in almost every industry since ChatGPT debuted in November 2022.
As with those in other industries, creative agencies want AI—but they want it on their own terms. At Cannes, many of the world’s biggest agencies—WPP, Publicis, Havas, and Dentsu—announced plans to use AI. Some of these agencies say AI is already helping them to mock up campaigns and storyboard spots. Some are using AI models to test ads—predicting their likely impact on different audience segments. And some are using the technology to create full campaigns.
But agencies’ race towards AI is driven as much by fear as it is by opportunity. AI may let brands bypass them entirely and use AI to design their own campaigns. At the very least, they worry AI will challenge a business model in which the agencies traditionally charged based on people hours. In a world where generative AI supercharges productivity, agency pricing models that were based on the human labor required for a campaign will be increasingly untenable, accelerating a trend towards performance-based pricing that is already well underway.
And, of course, everyone working at agencies is worried about what AI will mean for their own jobs. At Cannes, that anxiety was evident just beneath the surface. At a breakfast event Fortune cohosted with Meta, a number of marketing executives asked me privately what I thought AI’s impact on jobs would be—both in the ad industry and more broadly. Ahead of the festival, advertising agency Publicis released a video meant to be—according to Publicis CEO Arthur Sadoun—a “not so serious” take on creative agencies’ AI obsession. The video poked fun at “old Mad Men pitching AI,” offering assurances that “our people will be fine,” but that soon switched to “here a cut, there a cut, everyone gets cut.” The video ended with the tagline “Imagine if we all took the BS out of AI?” I thought the video was mildly amusing—and landed some valid points with its snark. But, according to an article in trade publication Campaign, the ad cut a little too close to the bone for some industry executives.
With AI, agencies are finding themselves, not for the first time, in the “frenemies” zone with tech companies—both partnering with them and heavily dependent on them as a marketing channel and yet potentially competing with them too. All along the beach at Cannes, Big Tech—companies such as Google, Meta, Snap, and TikTok—used their big tents to showcase their latest ad tech, touting new AI-powered tools for ad development and placement.
At the Fortune-Meta breakfast event, Erik Hawkins, Meta’s vice president for global partners, pointed out that brands that use Meta’s AI tools on average see a 22% increase in return on every marketing dollar they spend. Many of Meta’s AI tools are simply back-end processes and allow for more precision targeting of ads. But some of them can also be used to generate a wider, more diverse range of marketing content too. Hawkins said that as these tools gain wider use, making sure they are used responsibly is becoming more important, and that Meta had joined dozens of other big companies, academic institutions, and nonprofits in forming the AI Alliance that is supporting open research and open science using AI.
At Cannes, many brands were divided on the use of generative AI to produce campaign content. They’ve watched as some companies such as Levi’s and Victoria’s Secret have faced consumer backlash after using AI to create marketing imagery. In both cases, the campaigns were meant to portray the brands as diverse and body-positive. But they faced objections that the use of generative AI belied that message because it deprived real human body-positive and minority models of work. Some brands, including H&M, L’Oreal, and Lego, have barred the use of generative AI in marketing materials. But others such as candymaker Mars and cosmetics giant Estée Lauder are jumping in with both feet—using generative AI technology to help hone marketing content or brainstorm marketing concepts.
My own prediction is that brand hesitancy over generative AI will crumble over the next few years, as those experimenting with the technology demonstrate its ROI and figure out how to sidestep the ethical pitfalls.
But I couldn’t help thinking at Cannes this year that the ad industry has still not fully internalized the magnitude of the platform shift that is about to hit them. If AI chatbots and assistants become our primary means of accessing the internet, and of conducting e-commerce (which is likely to happen as AI models improve), how brands reach consumers is going to change profoundly. It won’t be about reaching consumers online anymore, because they won’t be there—their AI assistants will be there in their stead. Already publishers are talking about licensing deals with companies such as OpenAI and Google, that will see their content used more prominently in chatbot responses. Both tech companies have broached the idea of displaying ads against chatbot responses. But if people just send their AI agents out to do their shopping, booking, and buying for them, brands may find their direct access to the consumer diminished and attenuated.
At Cannes, the tide was turning. But it wasn’t clear those partying in the beach tents and on the yachts had noticed.
With that, here’s more AI news.
Jeremy Kahn
jeremy.kahn@fortune.com
@jeremyakahn
Today’s news, research, calendar, Fortune on AI, and Brain Food sections were curated by Fortune’s Sharon Goldman.
**Before we get to the news, a quick reminder to preorder my forthcoming book Mastering AI: A Survival Guide to Our Superpowered Future. It is being published by Simon & Schuster in the U.S. on July 9 and in the U.K. by Bedford Square Publishers on Aug. 1. You can preorder the U.S. edition here and the U.K. edition here.
AI IN THE NEWS
McDonald’s deep-sixes drive-thru AI experiment but says AI order-taking will still happen. In a sign that it remains stubbornly difficult to add AI to products millions of consumers depend on, McDonald’s ended a partnership with IBM in which it experimented with voice AI order-taking at the drive-thru. According to Restaurant Business, More than 100 restaurants had been part of the pilot. But the fast-food giant did not close the drive-thru window on future voice AI ordering options, saying it would evaluate other solutions by the end of the year.
The International Monetary Fund warns of labor disruptions and rising inequality due to generative AI. According to the Financial Times, the International Monetary Fund said it had “profound concerns” about massive labor disruptions and rising inequality as societies move towards generative AI, and it urged governments to do more to protect their economies. In a report published on Monday, the fund said countries should act such as improving unemployment insurance, warning that, unlike past disruptive technologies, AI could lead to job losses in higher-skilled occupations.
TikTok introduces customizable digital avatars and language dubbing features to "add a human touch." The Verge reported that TikTok is debuting new generative AI options for brands and content creators to spin up customizable digital avatars and language dubbing features. But far from implying that social media influencers would outsource their posts to their avatar counterparts, the company said the new tools are meant to expand language options and intended to “add a human touch” to content that would not have included human models or presenters.
EYE ON AI RESEARCH
AI-generated video is about to exit the 'silent film' era. All those cool AI-generated videos out there from OpenAI, Runway, and other startups have one thing in common: They are typically silent. So how can you add just the right sound effects and music to your creations? Google DeepMind shared its progress on video-to-audio AI that can generate soundscapes synchronized to on-screen action. Researchers said the technology, which uses video pixels and text prompts, will pair with video generation models to create shots with a dramatic score, realistic sound effects, or dialogue that matches the characters and tone of a video. It can also generate soundtracks from traditional footage including archival material and silent films.
FORTUNE ON AI
Canada-based Waabi raises $200 million, touts plan to get gen AI driverless trucks on the road in 2025 —by Sharon Goldman
Companies crave fresh data to train AI models. This startup’s recipe? Data made from scratch—by AI —by Sharon Goldman
AI CALENDAR
June 25-27: 2024 IEEE Conference on Artificial Intelligence in Singapore
July 15-17: Fortune Brainstorm Tech in Park City, Utah (register here)
July 30-31: Fortune Brainstorm AI Singapore (register here)
Aug. 12-14: Ai4 2024 in Las Vegas
BRAIN FOOD
What a headless flamingo says about authenticity in the generative AI age. In times past, a photo of a seemingly headless flamingo would likely have triggered feelings of wonder, awe, and curiosity. Now, many of us might simply think—AI. But in a turnabout twist, a photo of a beheaded puffy pink bird on long, pink legs was honored in a photography contest’s AI category, but promptly disqualified when the photographer revealed that the photo was real—the feathered friend was simply having a scratch. The Washington Post spoke to photographer Miles Astray, who said he captured the photo two years ago in Aruba. “My goal was to show that nature is just so fantastic and creative, and I don’t think any machine can beat that,” Astray said. “But, on the other hand, AI imagery has advanced to a point where it’s indistinguishable from real photography. So where does that leave us? What are the implications and the pitfalls of that? I think that is a very important conversation that we need to be having right now.”