Key Takeaways:
- AI adoption strategy appears highly sensitive to context: while delivering uncontroversial value as “invisible infrastructure” for back-end efficiency over the holidays, its deployment in emotionally-driven, visible creative work – especially when the output is bad – triggers a visceral consumer response.
- A new decline in dedicated metaverse investment following Meta’s push suggests that consumer demand for digital identity and virtual goods is platform-native, not destination-driven. These core behaviors continue to flourish organically within established ecosystems like Fortnite and Roblox, indicating that adoption prioritizes existing user flow over new world destinations.
- Fortnite’s model remains fashion’s blueprint for digital goods. Collaborations like the SKIMS-inspired outfits continue to demonstrate that digital fashion works best when it integrates seamlessly into existing user experiences, rather than demanding new environments
It’s always hard to get a read on how culture, en masse, feels about technology. But that’s especially true about the outward-facing applications of AI, where sentiment is, at this point, irreversibly polarised. The same tools are now appearing everywhere at once, with very different results – from creators working to slap “No AI was used” labels on their digital work, to others who see the potential as pure democratisation and market opportunity.
Case in point: take the divergent cases of two of the “words of the year” for 2025. “Vibe coding” made the top of the list for Collins Dictionary, putting a largely positive spin on the idea of people creating their own software instead of being beholden to what big technology companies create. On the opposite end of the definitional spectrum, “slop” (the pejorative du jour used to describe high volume, low quality output in essentially any form) was at the top of the peak for Merriam-Webster.
Whatever people feel about AI, it’s clear they feel it strongly. But this is, of course, only applicable to consumer-facing and end user applications. Planning, forecasting, and product discovery are all areas where machine-led systems are doing real work, improving speed and scale without demanding much attention. Black Friday offered a good example. AI was often credited with lifting digital performance, and very few people seemed particularly bothered by that fact, because back office execution is something consumers expect without wanting to think too much about.
As evidence of the difference between outward-facing and back-office AI use: consider the backlash against the McDonald’s and Coca-Cola AI Christmas campaigns. Christmas advertising remains one of the few surviving cultural slots where brands are still expected to perform sincerity rather than optimisation, and where audiences are unusually sensitive to signals of effort, care, and human presence. The pushback did focus in some aspects on the technical quality of the work (which, to anyone actually up speed with what’s possible in generative video was pretty far behind the curve), but most were grounded in the feeling that automation had wandered into one of the most sacred and sentimental spaces in advertising, if there is such a thing.
Coke’s Christmas advertising, for instance, is usually built by extremely capable CGI and animation teams, and audiences have assigned it a special seasonal position on that basis. When that level of craft is swapped out for something that feels, to be charitable, like a circular pastiche of itself, it’s hard for the people on the other end of the value exchange not to feel cheated. By attempting to echo the visual cues of past campaigns filmed with real actors and in real locations but using AI, it ended up highlighting the gap rather than closing it.
Where AI work seems to land more favourably is when it stops trying to compete on that ground altogether. The Valentino X Vans project via Harmony Korine’s EDGLORD didn’t look like it was attempting to recreate the language of traditional advertising like for like, or smooth over the oddness of the technology. It felt, funnily enough, like it may have leaned into it, though we hasten to add that’s an in-house theory as opposed to confirmed fact. The abstraction, the slightly splintered narrative, the sense that things weren’t quite lining up all became part of the piece, rather than something to apologise for.
But if selling AI to the mass public is a fraught path, spare a thought for the metaverse huckster. For years, Mark Zuckerberg enthusiastically framed the concept as something people were just about to step into, a connected digital world ready for habitation. Of course it was anything but – even if Meta committed to that idea deeply enough to rename itself, and then continued investing as the promised adoption failed to appear. This month’s reports of significant funding cuts to metaverse initiatives within the company (alongside a clearer push toward, you guessed it, AI) read like an acknowledgement that the original bet was a weird one.
What’s slightly awkward in hindsight is that many of the behaviours the metaverse was meant to unlock never actually failed. They just didn’t happen inside the structure that was supposedly being built for them. Digital identity, virtual goods, and branded self-expression have all carried on quite happily elsewhere, without the need for a newly declared destination or a lot of future-facing language. Video games have been absorbing these behaviours for years without insisting they represent a new phase of anything, as we’ve documented here at The Interline throughout our existence.
That’s why the Fortnite collaboration with Kim Kardashian this week fits so easily into the picture, with none of the backlash that would have been associated with an AI campaign, or a Horizon Worlds partnership. SKIMS-inspired outfits dropped into the game (that Kardashian reportedly had a real hand in how they looked) and, by all accounts, did a good job of both representing the brand (SKIMS and Kim herself) as well as fitting into the established Fortnite aesthetic and the existing content ecosystem.
Digital garments in these spaces don’t need to argue for their authenticity or permanence, which is where earlier digital fashion experiments often became overburdened. They get worn, shared, and eventually replaced by the next culturally hot drop Epic has lined up. Much of what struggled in metaverse-first projects wasn’t the idea of digital fashion itself, but the expectation that people would adopt new behaviours simply because a new world had been declared important. And the same twisted logic is behind some of these outward-facing AI content campaigns: the idea that people will warm to something because the creation of it was a novelty or an experiment, even if all it’s doing is remixing things they’ve seen before.
As our DPC Report 2026, releasing this week, gets into, the integration of AI into existing technologies, and into the content ecosystem is a complex thing, with a lot of valid use cases and arguably just as many short-sighted and invalid ones. For fashion and beauty brands, the trick of separating the two might still be technically difficult, but culturally it seems, to us, easier than ever: cultural cachet is the only game in town, whether you’re generating your media or creating it the old-fashioned way.
