What changed and what did not
Snap said it had changed its recommendation systems to promote only videos made by real people. This is not a ban: creators may still use AI tools to enhance or edit their content. In a blog post the company said it wanted Spotlight to remain a place where people discover authentic creativity from real people. The announcement gives no date for when the change takes effect and shares no detail about creator payouts.[1]
The resource circulating in Spotlight is distribution. Changing what the recommender promotes changes who gets seen; that is a measurable effect, and the affected parties are creators. Whether it changes who earns is a separate question, and not one the announcement answers. The link between being seen and being paid runs through payout rules, and whether those rules were touched is not stated.[1]
Who draws the line
The decision operates at the line between fully AI-generated content and AI-enhanced content. Snap defines that line, Snap applies it, and creators carry the result. The announcement gives no threshold, no account of how the classification is made and no appeal route for a misclassified video. A creator whose reach falls has nothing that would distinguish the fall from ordinary fluctuation in recommendations.[1]
That does not show the change works against creators. Content pulled back in recommendations may keep circulating on other surfaces of the app and may not change total views measurably; that may also have been the case after the comparable step in April 2026. Without payout data the direction of the effect cannot be established either way. What can be said is that the party carrying the consequences of the decision is not the party making it.[1]
What could be stated as plainly as the policy
Snap can state the policy in a single sentence. Two quantities could be stated with the same clarity: what share of Spotlight recommendations goes to human-shot video, and how creator payouts changed over the same period. Both are figures the company already measures, and both are of the kind that would show whom the change works in favour of. The same two figures can be asked of the comparable changes at LinkedIn, Meta, YouTube and Substack.[1]
If by 31 December 2026 Snap discloses only the reach side of those two figures, the change has stayed on the distribution side; if payout data follows, whom the effect is credited to becomes measurable. Announcing an effective date is a signal in itself, because a policy without a date cannot be measured.[1]