Unique Research · Outlook · January 3, 2025
Over the past year, tech giants and startups raced to ship generative AI tools for users and advertisers. From AI images in programmatic ads to a growing wave of AI-made TV spots, brands are testing cross-platform creative. Late 2024 buzz focused on wider access and better output from models such as OpenAI’s Sora, Amazon’s Nova, and Google’s Veo.
The tech impresses — and polarizes. Some see a powerful new creative form; others call it disappointing “AI slop”; others fear IP theft and job loss. Which view wins will matter.
In the year ahead, tools from TikTok, Meta, Amazon, and peers may spread further among users, creators, and advertisers. Better quality and lower cost could spur mass AI content. How copyright rules and lawsuits shape 2025 innovation and adoption remains unclear.
Pika recently shipped AI video model Pika 2. Lindsay Brillson, Pika’s brand and content lead, says more brands use Pika for organic social and ads: “It’s fun, gets attention, and fits the platform. We can truly control scene building blocks — characters, backgrounds, wardrobe, props, products — instead of leaving it to chance.”
Pinterest, Meta, and Snap spent the year on AI ad generation and new user features. Accenture’s Kevan Yalowitz notes creator and enterprise content tools are converging — which may drive adoption — and is bullish on more personalized, dynamic ads: “For the first time they can tailor ads to a single user, even a single search. That idea became real on most platforms this year.”
VaynerMedia’s Jon Morgenstern expects SMBs to keep adopting platform AI content and ad tools — citing TikTok Symphony Creative Studio, where users paste a URL to create AI content. Large brands hesitate over privacy, copyright, and IP, which may advantage smaller ones: “Niche brands will use these tools immediately. It disproportionately levels the field — challengers fly under the radar and aren’t sued the way big brands might be.”
DAIVID CGO Barney Worfolk-Smith says generative AI’s impact on creative assets is not fully felt. DAIVID uses AI to measure how ads hold attention and move emotion. He sees potential in predictive analytics and synthetic audiences, but thinks personalization-at-scale claims are oversold and can hurt quality — though better output and lower cost may change that.
“We compared emotion from hand-made vs AI-made Coke ads,” he said. “Warmth is a standard Christmas-ad emotion; in the AI version it dropped about a third. We basically attribute that to the uncanny valley.”
As platforms give creators AI tools, questions remain about compensating creators whose work trains models others then use. Gartner’s Andrew Frank points to Bria, Getty Images, and Adobe efforts to pay contributors based on influence on AI outputs.
“We’re at a peak of anxiety — not only politics, but because things change so fast,” Frank said. “People still feel future shock from this year’s AI and other tech. It feels rapid — and a bit out of control.”
Whether 2025 brings more control is open. What is not: tech providers, adopters, and humans will keep inventing ways to build, deploy, and test new AI creative.
— END —