Image: relevantmagazine.com · rights & removal

It’s Not Just You: Instagram Has Been Pushing Creed’s ‘Higher’ on Everyone’s Stories But Doesn’t Know Why

Reporting by Relevant MagazineRead the original at relevantmagazine.com

Executive Summary

A company spokesperson stated that an investigation into song recommendations on Instagram found that Creed’s “Higher” was recommended to more people than intended, and the engineering team is working on a fix. One source noted that the song is always recommended to them personally. Users reported noticing this phenomenon over several weeks, with some suggesting they were uncovered something significant. Instagram's system is designed to recommend content based on user history and selected media, yet users observed recommendations for songs like “Higher” appearing regardless of the visual content. Another song suggestion, Soulja Boy’s “Turn My Swag On,” was recommended, which Instagram explained reflects actual popularity rather than a bug. The situation highlights a discrepancy between intended algorithmic function and observed user experience regarding music suggestions on Stories.

Facts Only

* A company spokesperson told CNN that the team investigated song recommendations and found Creed’s “Higher” was recommended to more people than intended.
* The company stated that its engineering team is working on a fix for the issue.
* One user posted on X in August suggesting Creed was paying META to suggest their song.
* Another user posted in September stating they discovered the issue affecting others.
* A third user referenced receiving suggestions like “Higher” with unrelated content, such as a dinner plate or a Caesar salad.
* Instagram is supposed to select songs based on signals like past usage and photo/video selection.
* A communications specialist noted that recommendations occurred when posting various content, including photos of a cat or a dinner plate.
* The company claimed the music suggestion was a mistake and was being fixed.
* Soulja Boy’s “Turn My Swag On” suggestions were explained as reflecting actual popularity.

Full Take

The narrative balances an official admission of error regarding one specific recommendation with user-observed experiences suggesting broader systemic issues with the recommendation algorithm. The tension lies between the stated system logic—that recommendations should be contextually relevant to media selection—and the observed reality where seemingly irrelevant popular music is suggested. This discrepancy opens up questions about the weight given to historical preference versus current content context in the Instagram algorithm, and whether mass-scale adjustments are prioritized over individual contextual relevance. The differential treatment of Creed versus Soulja Boy suggests that algorithmic function may be unevenly applied or prioritize different types of data differently depending on the track's perceived popularity or embedded weight within the system. It prompts inquiry into what constitutes "intended" user experience and who benefits from the maintenance or modification of these recommendation systems when they diverge from explicit design goals.

From the original · Relevant Magazine

If you posted a photo of your dog, a dinner plate or a caesar salad to your Instagram Story lately and the app suggested Creed’s “Higher” to go with it, you weren’t imagining things. A company spokesperson told CNN the team “looked into this and found that this song is being recommended to more people than intended.”

Read the full story at relevantmagazine.com

Sentinel — provisional

Raw score, 0 to 1 (uncalibrated)0.35

No strong signs of machine writing were found in the source article. Provisional estimate, not a finding that a person wrote it.

The model's own note (unverified)

The text reads like journalistic reporting attempting to synthesize specific anecdotes with official responses regarding an algorithm feature, lacking the hyper-polished symmetry often associated with pure machine generation.

This looks only at the wording of the original source article, not at this page's AI-written sections. A small local AI model made this estimate. It has not been checked against known human and machine texts, so treat it as provisional. It cannot show who wrote an article.

What the model noticed in the source article

Sentence length variance is moderate; the text shifts between direct reporting and anecdotal voice.

The narrative flows logically from an anecdote to a claim to a counter-claim, reflecting a somewhat chaotic but intentional framing.

Relatively direct attribution is present (spokesperson quotes) mixed with subjective user claims.

The text relies heavily on reported, unverified user claims, which inherently introduces a degree of narrative uncertainty regarding the core premise.

Signs the model read as against machine writing

The inclusion of highly specific, emotionally charged anecdotal evidence ('dinner plate or my cat or whatever') suggests an attempt to ground a larger claim in relatable experience rather than pure data presentation.

The shift between corporate response and user suspicion feels like the necessary friction of reporting on an ongoing investigation.