TRENDING: KOREAN BASEBALL & BUILDING A JUNK DRAWER
KOREAN BASEBALL AI TREND
TREND
The biggest format of the month, full stop. Creators upload a single selfie and use AI tools (most notably Kling AI, which has shot to #1 on the App Store in 42 countries off the back of this) to generate a 4 to 5 second clip that looks exactly like a KBO (Korean Baseball Organisation) broadcast camera randomly cut to them in the stands: stadium lights, blurred crowd, lower-third graphic, the slight zoom of a real TV camera finding a face. It detonated globally in a tight 10-day window after a single "the average Korean woman" clip racked up 15 million views on X in early May, and the reason it lands is the uncanny realism. The output is indistinguishable from a real broadcast frame, which makes the gag genuinely funny every time.
The backlash and the parody wave
There's a strong critical reaction running in parallel. Korean media flagged early that the original clip wasn't labelled as AI, sparking a wider conversation about audiences' ability to tell real from fake content. Creators and commentators have also called out the trend's beauty-standards problem (almost every original variant featured a young, slim woman in a white top), and K-pop fans have pushed back on AI edits of idols like Jungkook being placed in stadium crowds, calling them invasive and creepy.
BUILDING A JUNK DRAWER
@middleclassfancy’s content of "building the perfect junk drawer" Reel is the latest example of a format brands and creators are leaning into hard in 2026: episodic, recurring series with a consistent character, premise, and pay-off, in this case, a deadpan suburban-dad voice walking through a perfectly curated junk drawer ("the loose batteries go here, the takeaway menus from 2018 go here, the Allen key from an IKEA bookshelf you no longer own goes here").
The format works because it's instantly familiar, hyper-specific to a kind of household everyone recognises, and it's part of a wider 2026 shift- Sprout's Q1 Pulse found 20% of consumers (27% of Gen Z) want more episodic series content from brands, and audiences are following recurring characters far more reliably than they're following brand handles.
BECOMING A X (BADLY)
The format of the week on TikTok: "Day 1 of trying to become a [DJ/sports commentator / stand-up comedian / radio host / opera singer]," with the creator then performing the role catastrophically badly on purpose, but the same scaffolding is being applied to commentary, podcasting, voiceover, cooking, anything with a recognisable "professional" sound. The point isn't to be funny in a polished way; it's to be bad in a way that's genuinely irritating, because the engine of the format is ragebait.
Commenters can't help themselves: explain why it's wrong, tag friends who'd hate it, or argue with each other about whether the creator knows it's bad. The algorithm reads that flood of comments and saves and quote-shares as high engagement and pushes the video harder. It's a textbook example of TikTok's current incentive structure: anger equals reach, and the "Day 1" framing gives the creator plausible deniability while the badness does the work.
