Millions of consumers now ask AI what to buy, where to travel, how to manage their health, and who to trust.
But here’s the question most brands haven’t asked yet: when multicultural consumers ask AI for guidance, does your brand show up at all? If it isn’t, the reason may not be AI bias—it’s that the content ecosystem AI learns from has a gap.
Early research is already revealing what is at stake. A recent study analyzing AI recommendations in the GLP-1 nutrition category found that the brands winning AI recommendations were not the most recognized; they were the most citable. Brands with clear narratives and credible coverage in authoritative sources consistently outperform household names with larger budgets.
A separate analysis of airline recommendations across Latin American markets revealed a similar pattern. American Airlines lost eight points in AI visibility the moment a traveler asked the same question in Spanish instead of English. Same route. Same traveler. Completely different answers.
These findings point to something bigger: a problem we call the Cultural Visibility Gap.
The Cultural Visibility Gap describes the difference between the audiences that brands believe they reach, and the audiences for whom those brands appear in AI-generated recommendations.
Language alone can change who AI recommends brands to, and who it leaves out. But language is just one dimension of culture. AI doesn’t learn from the world as it is lived; it learns from the record the internet has left behind. For decades, multicultural communities have been underrepresented or simply absent from mainstream media and digital content ecosystems. When AI models absorb that material, they’re not making a judgment call—they’re reflecting what exists in the record. For multicultural audiences, that record is thinner. Not because AI is biased, but because the content ecosystem itself has a gap. And as AI gets better at personalizing responses for individual users—learning who you are through behavioral signals rather than explicit prompts—the absence of in-culture content in that record won’t shrink the gap, it will deepen it. Brands that are not visible today will be even harder to find tomorrow.
So, we looked more closely
The gap is documented. What’s less understood is what it looks like for brands specifically, which ones show up, which ones disappear, and why. So we took a closer look at a high-stakes category: first-time homebuyers, one of the most important financial decisions a family can make, and one where multicultural consumers represent a significant and growing share of buyers.
We observed how AI responded to the same homebuyer questions across multiple platforms, tracking every brand mentioned, every program recommended, and every shift in tone, with and without cultural and language context introduced.
The results were telling. Without any cultural context, AI responded like a confident advisor across the board: national lenders, standard programs, specific comparisons.
When language and cultural context entered the picture, the recommendations shifted. Different lenders appeared. Community-specific programs surfaced, including state and city-level down payment assistance programs that never appeared in the general market response. On some platforms, the response acknowledged cultural dynamics directly. On others, brand-name lenders largely disappeared, and the response redirected toward the buyer’s financial readiness rather than engaging with the market itself.
The point isn’t that AI got it wrong. The point is that brand visibility changed, significantly, based on cultural context. For the general market buyer, major national lenders showed up consistently. For culturally specific buyers, that landscape shifted, and for some platforms it narrowed considerably.
This is not a glitch. It is a content problem. Mainstream brands have deep earned media in mainstream outlets. They haven’t built equivalent earned media within culturally specific communities. So when AI is asked to recommend something to those audiences, it’s drawing from a thinner record. This is a brand visibility problem—but only because it’s a content ecosystem problem. And most companies don’t know they have it.
The Earned Media Connection Most Brands Are Missing
Understanding why the gap exists also points to how brands can close it. AI tools learn from what exists across the web: articles, features, mentions, and stories. Earned media is part of that record. Brands aren’t fighting AI bias when they build earned media in culturally specific communities—they’re enriching the training data AI learns from.
PR placements in culturally relevant outlets such as Essence, Remezcla, Axios Latino, community news sites, and ethnic media become part of the authoritative record AI draws from when making recommendations to those audiences. This isn’t about gaming algorithms or keyword optimization. It’s about authentic presence in communities where your brand actually operates.
A brand can be well optimized for general market AI visibility and still be absent from the results that matter most to Hispanic, Black, Asian, and New American audiences. Where a brand earns coverage matters. The outlets themselves shape the audiences for whom that brand becomes visible—and the content ecosystem AI learns from.
The framing of earned coverage matters as well. Content that truly reflects cultural context, uses the language people actually speak, and showcases genuine community ties becomes part of the content ecosystem AI draws from. It helps make a brand findable, credible, and relevant to multicultural audiences in AI-generated answers. Repurposed general market messaging doesn’t do that work.
Earned media has always been about trust. In the age of AI, it is also about shaping the content ecosystem that AI learns from.
The Window Is Open, But It Won’t Stay That Way
What makes this moment unusual is that the gap is visible but largely unclaimed. Most brands are not asking these questions yet, and most agencies are not testing how AI responds to different audiences. But as AI gets smarter at inferring who users are—and personalizing accordingly—this gap will matter more, not less.
AI recommendation systems reward consistency, citation depth, and content authority. Every culturally relevant article placed, every in-language resource published, and every earned mention in a trusted community outlet becomes part of the record AI draws from, and that record compounds over time. The brands that start building earned media within these communities now will be part of the authoritative record AI learns to recommend from tomorrow. The ones that don’t will find themselves increasingly absent.
Ready to build an earned media engine that reaches diverse audiences in the age of AI?
