On July 14, 2026, WARC published a report in collaboration with TikTok and LIONS Advisory that puts an exact number on something the advertising industry had been sensing for the past two years: generative AI has multiplied the volume of advertising content without a proportional improvement in quality. 88% of the 400 marketers surveyed in the United Kingdom, United States, Australia, and Brazil say AI has increased their creative output, but only 45% say it has significantly improved quality. The gap between those two figures is, essentially, the real problem facing any brand that currently uses AI to produce TikTok ads.
The report, titled The New Creative Advantage: How Community Signals Are Reshaping Creative Success in the Age of AI, does not conclude that AI is the problem. It concludes that the problem is what information you give AI before asking it to generate something. This article explains the study’s key findings, why most brands are still feeding their AI tools with demographic data they themselves no longer believe is effective, and what you should change in your creative process for TikTok Ads as a result.
What Did the TikTok, WARC, and LIONS Report Find?
The report found that 90% of marketers already consider AI an established part of their creative toolkit, but most are still feeding it the same type of information they used before generative AI existed. The research combined a survey of 400 marketers directly responsible for creative and content decisions, conducted in May 2026, with interviews with marketers and industry experts, as well as a review of global WARC and TikTok data.
The study’s most uncomfortable finding is this: 67% of marketers say demographic data remains the most common input they use to instruct AI, even though 59% acknowledge, according to WARC’s own Marketer’s Toolkit 2026, that traditional demographic targeting is no longer effective. In other words, most brands know their targeting method is outdated, yet they continue using it to feed the most powerful marketing tool they have had in years.
In contrast, only 17% of marketers say they systematically incorporate community or real-audience signals into their generative AI workflows. According to the report itself, this 17% is the group gaining a real creative advantage over the rest—not because they use better AI tools, but because they give them better instructions.
Why Doesn’t More AI Content Mean Better Quality?
More volume does not mean better quality because an AI model can only work with the information it receives, and a generic demographic label, a recycled brief from previous quarters, or a list of already-known trends gives the model nothing new to interpret. Generation speed does not solve a poor-information problem; it simply repeats it faster and in more variations.
This distinction is important because the report does not say that AI-generated content is inherently worse. It cites evidence from System1 tests in which AI-generated ads scored significantly above the global advertising average, as well as a large-scale analysis by Columbia Business School that found AI-generated display ads performed competitively in the real market. The problem is not the technology; it is the input provided before asking it to produce something.
This section stands on its own: if your team already uses AI to generate TikTok ad variations and you feel that you are producing more but not necessarily better content, the most likely cause, according to this report, is not the tool you use. It is that you are still describing the same generic audience to that tool as always.
What Are Community Signals and Why Do They Outperform Demographic Data?
Community signals are the information generated by a real audience when it searches, comments, shares, remixes, creates, and buys within the same participatory environment, rather than the static label describing who that audience is supposedly made up of. A demographic data point tells you that your buyer is between 25 and 34 years old and lives in a large city. A community signal tells you what specific format that same audience is remixing this week, what objection keeps appearing in your competitors’ comments, and what language real people use to describe the problem your product solves.
Lexi Wolf, Director of Thought Leadership at LIONS Advisory, summarizes the report’s central contradiction directly: belief in the effectiveness of demographic targeting is already declining among marketers themselves, yet that same group continues to use demographics as the number-one input for instructing AI. Marcos Angelides, General Manager of L’Oréal Lab and Head of AI Operations at Publicis Media, puts it even more directly within the report: the advantage lies in the data you train or feed the model with, because AI is only as good as the data it receives.
Andy Yang, Global Head of Ads Creative and Brand at TikTok, describes the pattern that separates winning brands from those that do not: they are not the brands generating the most content; they are the ones learning fastest from the people they serve, a concept the report calls cultural intelligence. It is literally the same term we have already been using to describe the cultural intelligence work in Hispanic marketing carried out by Jorge Perez, founder of JP Director, and profiled by Newsweek in July 2026. It is not a coincidence: it is the same idea applied by two different sources, and both independently reach the same conclusion.
What Is the Intelligence Loop?
The Intelligence Loop is the framework proposed by the report to describe how the most effective brands continuously connect their audience with their creative process: audience participation generates signals, those signals reveal real demand, demand shapes creative, and that new creative generates more participation, closing the loop. It is a model of continuous learning, rather than a brief defined once a quarter and executed without review until the next one.
The practical implication of this framework is that it moves brands from reacting to trends after they have already exploded to using live signals to inform creative decisions while those trends are still taking shape. On participatory platforms such as TikTok, where audience behavior is visible in real time through comments, remixes, and search patterns, this cycle can be fed much more directly than on channels where the audience is more passive.
For this to work, however, someone has to actively collect those signals and translate them into concrete instructions for AI. The loop does not close simply because you have access to a platform with a high level of community activity; it closes when someone on your team turns that activity into a real input for the creative brief.
How Does This Apply to Your TikTok Ads in Practice?
In practice, this means replacing or supplementing the traditional demographic brief with a simple exercise before generating any ad variation with AI: review what people are actually commenting on, remixing, or asking about your product category this week—not last quarter. We have already covered in detail the principles that make an ad actually convert in 2026, and this report confirms with data something we were already applying there: the initial hook and the core message matter more than the number of variations you generate.
One concrete exercise is to measure the relationship between hook rate (how many people keep watching during the first few seconds) and the qualified click-through rate toward the next step of your funnel, rather than looking at either metric independently. A strong hook with low commercial intent can be entertaining without supporting the campaign’s actual objective, which is exactly the type of outcome produced by AI that is well-instructed in entertainment but poorly instructed in your business.
To illustrate this—not as real data, but as a typical scenario—imagine a brand generating fifteen variations of the same ad with AI using the same demographic brief as always, with all of them performing similarly and mediocrely. The same brand, feeding the AI three real community signals extracted from that week’s comments (a repeated objection, a format that is performing well, and a phrase the audience uses to describe the problem), generates only five variations, but two of them clearly outperform the previous fifteen. Volume went down, quality went up—exactly the pattern described by the report.
What Should You Do With Your Brand Now?
The first step is to audit what information you are currently giving any AI tool you use to generate TikTok creative, and honestly determine whether that input is a generic demographic label or a real, current community signal. If the answer is the former, you do not need a different AI tool—you need to change what you describe to it before asking it to generate something.
From there, build a simple, repeatable process: before each creative production cycle, spend time reviewing real comments, remixes, and questions from your audience on TikTok, and turn two or three of those observations into concrete instructions for the brief, instead of always starting from the same target-audience document. We have already covered the fundamentals of this type of strategy in our updated TikTok Ads agency guide for 2026, and this report confirms that this approach is not merely our editorial preference—it is a pattern measured across 400 real marketers.
At JP Director, we apply this same cultural intelligence logic described in the report, not as a platform feature that you simply turn on, but as a deliberate step before any creative production with AI: understand the real audience first, then generate. That is the difference between using AI to produce more of the same thing faster and using it to produce better versions of what your audience is already telling you it wants to see.
Frequently Asked Questions
Does the TikTok and WARC report say that generative AI does not work for advertising?
No. The report cites evidence that AI-generated ads can perform very well, including System1 tests in which AI-generated ads scored above the global advertising average, as well as a Columbia Business School analysis that found AI-generated display ads performed competitively in the real market. The report’s central finding is not that AI fails; it is that the quality of the output depends directly on the quality of the information it receives before generating.
What are community signals in the context of TikTok Ads?
Community signals are the information generated by a real audience through its visible behavior on the platform: what it comments on, remixes, shares, and asks about a product category at a specific point in time. Unlike static demographic data, a community signal changes constantly and reflects what the audience actually thinks and needs right now, rather than a generic profile defined months in advance.
How many marketers actually use community signals to guide their AI?
According to the WARC, TikTok, and LIONS Advisory report, only 17% of surveyed marketers say they systematically incorporate community or real-audience signals into their generative AI workflows, while 67% continue using traditional demographic data as their primary input, even though most recognize that this type of targeting is no longer as effective as it once was.
What is the Intelligence Loop?
It is the framework proposed by the report to describe how the most effective brands continuously connect their audience with their creative process: audience participation generates signals, those signals reveal demand, demand shapes new creative, and that creative generates more participation, closing the cycle. The goal of the framework is to move brands from reacting to already-established trends toward using real-time signals to anticipate creative decisions while those trends are still taking shape.
Last updated: July 2026. The full report, The New Creative Advantage, is available through WARC and TikTok for Business; verify the complete findings directly from the source before basing budget decisions on specific figures.
