MediaScience announced on March 18, 2026 that it has rolled out a new AI‑driven capability called Creative Twin™. The feature, built on the company’s MediaPET.ai platform, claims to generate exact digital replicas of existing advertisements and then allow marketers to tweak individual components—such as talent, visuals, or copy—to measure their isolated impact on audiences. The launch was slated for presentation at the Advertising Research Foundation’s Audience x Science conference, where founder and chief executive Duane Varan is expected to demonstrate the technology.
A synthetic copy of any ad, indistinguishable from the original
Creative Twin leverages proprietary machine‑learning models to reconstruct a video or display ad down to the pixel level. According to MediaScience, the resulting AI‑generated version is “indistinguishable from the original creative.” The process occurs entirely within the MediaPET.ai environment, a spin‑off focused on AI‑based media analysis, and does not require additional filming or post‑production work.
Independent validation backs the claim
To substantiate its promise, MediaScience partnered with the Ehrenberg‑Bass Institute—an academic hub for marketing science—and conducted a controlled study involving 812 U.S. respondents. Participants were shown either the original advertisement or its AI‑crafted twin, without being told which was which. The test found that viewers could not reliably tell the two apart, suggesting that the synthetic version maintains the visual fidelity and emotional cues of the source material.
Beyond visual similarity, the study demonstrated that once an ad is cloned, each element can be altered in isolation. This capability opens the door to measuring the precise contribution of a single variable—such as a celebrity endorsement or a specific visual motif—on key performance indicators like brand recall, attitude, and purchase intent.
Why granular creative testing matters now
Traditional ad testing often relies on creating multiple full‑scale versions of a campaign, a process that can be both time‑consuming and costly. By contrast, Creative Twin enables a “digital sandbox” where marketers can experiment with countless permutations without incurring additional production expenses. The technology aligns with the broader industry shift toward addressable advertising, where messages are tailored to specific audience segments in real time.
For example, a shampoo brand could produce a single high‑quality spot and then digitally swap a straight‑haired model for a curly‑haired one when targeting consumers who style their hair. In MediaScience’s internal test, the curly‑hair version outperformed the original on metrics tied to brand recognition, attitude, and choice—indicating a measurable lift in purchase likelihood.
Real‑world use cases illustrate flexibility
- Pet‑food marketing: A commercial featuring a Labrador Retriever can be automatically re‑rendered with a poodle or French bulldog, allowing brands to serve breed‑specific versions without reshooting.
- Coffee advertising: Brands can swap out a high‑profile celebrity for a generic model—or replace the celebrity altogether—to gauge the incremental value of star power.
In each instance, the AI‑modified asset retains the production quality of the original, preserving brand standards while enabling rapid, data‑backed iteration.
Potential ripple effects across the ad ecosystem
If adopted at scale, the ability to test creative elements in isolation could reshape budgeting decisions for agencies and brands alike. Production houses might see a reduction in demand for multiple shoot days, while media planners could allocate more of their spend toward data‑backed creative optimization. Moreover, the technology could accelerate the move toward hyper‑personalized video experiences in programmatic environments, where a single master asset can be dynamically altered to suit diverse audience profiles.
However, the rise of synthetic media also raises questions about authenticity and disclosure. While the press release does not address regulatory considerations, industry observers have noted that transparent labeling of AI‑generated content may become a compliance requirement as synthetic media proliferates.
Executive perspective
“This represents a fundamental shift in how advertising creative can be evaluated and optimized,” said Duane Varan, CEO of MediaScience and MediaPET. “For the first time, researchers can isolate and measure the contribution of individual creative elements within an advertisement, providing marketers with unprecedented clarity about what truly drives effectiveness. And they can now properly optimize and personalize ads without compromising on production quality.”
Industry context: AI and synthetic media in advertising
Creative Twin arrives amid a surge of AI tools that generate or modify visual content—from deep‑fake videos to text‑to‑image generators. While many of these solutions focus on content creation, MediaScience’s offering emphasizes measurement and optimization, positioning it as a bridge between creative production and performance analytics.
Competitors in the ad‑tech space have begun experimenting with AI‑assisted creative generation, but few have publicly demonstrated the ability to conduct rigorous, statistically valid tests of individual creative variables. If MediaScience’s validation holds up in broader market deployments, the company could claim a first‑mover advantage in the niche of AI‑enabled creative experimentation.
Limitations and next steps
The initial study involved a relatively modest sample of 812 U.S. respondents, and the tests were conducted in collaboration with a single academic institute. Wider adoption will require additional validation across diverse markets, media formats, and cultural contexts. Moreover, integrating Creative Twin into existing ad‑ops workflows will demand coordination between creative teams, data analysts, and programmatic platforms—a logistical hurdle that could influence the speed of uptake.
MediaScience has directed interested parties to its websites, mediascience.com and mediapet.ai, for further details and potential pilot opportunities.
Get in touch with our Adtech experts
