What happened
On July 31, Business Insider reported that Snap said it will no longer recommend or reward fully AI-generated videos in Spotlight, the company's short-form video feed. The report said Snap is shifting its algorithm toward content made by real people, while still allowing creators to use Snap's AI tools to enhance or edit videos.
Snap's own recommendation eligibility guidance supports the policy direction. In its Quality section, Snap says its content ranking algorithm rewards authentic, human-made content over wholly AI-generated content created outside Snapchat, even when that AI-generated content carries transparency disclosures. The same guidance says AI-generated content made within Snapchat is eligible for recommendation and that Snap's tools include transparency indicators.
Snap's creator monetization policy gives the business side of the same idea. It says monetized accounts should demonstrate originality and authenticity, and that content using sophisticated editing or AI-based tools can be monetizable when it is original, entertaining, or informative, not misleading, and disclosed in the content or creator profile.
The line Snap is drawing
Snap's public policy allows AI tools, including editing and enhancement, when the work remains original and transparent.
Human-shot video can also be repetitive or misleading. Snap's monetization policy lists formulaic assembly, unoriginal reposting, and misleading setups as problems.
Record the contribution behind the clip
A label can tell viewers that AI was involved. It cannot tell them why the video exists, what the person knows, what evidence shaped the claim, what tradeoff the speaker made, or whether the final edit still represents the person whose name is on the account.
Snap's policy names creative effort, editorial judgment, personal contribution, and disclosure. Review each of those separately when assessing a clip.
Before publishing a clip, keep the original answer and transcript, preserve any caveat, note what AI changed, and have the speaker review the final claim.
Why recorded answers help
A research-guided video interview gives a team a stronger source than a prompt asking for a short-form video idea.
In an interview workflow, the person supplies the examples, vocabulary, uncertainty, and judgment. The editor can then turn that source into clips, posts, captions, or articles. AI can help prepare questions, organize the transcript, suggest cuts, draft captions, or clean up structure. But the public claim still starts with a human answer.
Short video compresses context. A viewer may see only thirty seconds, or encounter a caption without the full explanation. Keep enough of the answer for the claim to remain accurate.
The recording does not make the content automatically trustworthy. People can ramble, overstate, or perform on camera. The advantage is that the team has an inspectable source. A reviewer can check whether the clip preserved the speaker's meaning. The speaker can reject a line that sounds stronger than what they actually meant. The final asset has a path back to a real conversation.
A practical pre-publish check
Before publishing an AI-assisted short video, run a simple contribution check.
First, name the human contribution. Did the person add a firsthand example, specific analysis, original commentary, a useful demonstration, a critique, a decision rule, or a caveat? If the answer is only that the person selected a template, the asset is weak.
Second, name the source. Can the central claim be traced to a recorded answer, transcript line, source document, product decision, research note, customer-approved quote, or observed workflow? If not, the clip may be too detached from the expertise it is supposed to represent.
Third, name the AI role. Was AI used for research support, question preparation, transcript cleanup, clip suggestions, visual enhancement, caption drafting, or formatting? Keep that role specific. A specific role is easier to disclose and easier to review.
Fourth, check the edit against the source. Does the video preserve the speaker's meaning, or did the packaging turn a careful point into a stronger claim? Short-form editing often creates trust problems by removing the caveat that made the original answer responsible.
Finally, check whether the creator would say the same thing in a live conversation. If the answer is no, the clip is not ready, even if the disclosure is technically accurate.