Direct answer

It is a short, dialogue-driven meet-cute: two people in an elevator, one wearing headphones loud enough to leak The Smiths, the other recognizing the song. The joke is the misheard confession — “I love the Smiths” / “Sorry?” / “I said I love the Smiths.” TikTok turned it into a couple, friend and coworker template; AI generators turned it into a two-photo product in November 2026.

Make my video

What is the “I love the Smiths” trend?

A two-line elevator scene from 2009, a 2024 TikTok revival, and a 2026 template that lets two people cast themselves.

Published November 10, 2026

The scene, shot by shot

The source is the 2009 film (500) Days of Summer. Tom, wearing large headphones, is listening to The Smiths. Summer steps in, can hear the music leaking, and says she loves the band. Tom, not hearing her clearly, asks “Sorry?” She repeats it. The elevator opens; she leaves; he is left saying “Wow.”

The scene is about 25 seconds in the film. The TikTok cut is shorter: the glance, the headphone touch, the line, the exit. That is the shape every remake copies.

Why it came back in 2024

Searchlight’s official clip was posted to TikTok on 1 February 2024 and crossed 5 million views. Through 2024–2026, couples and friends reenacted it in real elevators, then in stairwells, dorm rooms and offices. The audio is the engine: the lines are short, the timing is forgiving, and the punchline is visual.

The 2026 template

By late 2026 the trend had a recognizable shot list and two named seats. The hashtag passed 12,000 posts and 300 million views; the search phrase “i love the smiths what i said i love the smiths” was still rising. AI tools launched two-photo versions within days of each other.

  • Two people, one elevator, one pair of headphones.
  • A fixed line reading: confession, confusion, repetition.
  • The held look and the reaction — no cut, no exit inside the 15 seconds.
  • The original sound, not a cover or a remake.

What AI changed

The first AI versions either cut the scene into five shots or generated a soft montage in one call. We went the other way: one 15-second take, 16:9, with the original sound, because the source is one continuous shot and every cut is a chance to lose an identity.