First-order effects

• A new TikTok account may encounter a feed dominated by low-quality AI-generated video before it has supplied enough viewing signals to personalize recommendations; the tested YouTube Shorts experience showed a materially lower share.
• TikTok faces sharper pressure to make its proposed AI-content controls and labeling effective in the For You feed, while YouTube must address the substantial AI-generated share still identified in Shorts.

Second-order effects

• Creators and advertisers seeking predictable audience environments may weigh feed quality and labeling clarity more heavily when choosing between TikTok and YouTube Shorts.
• Platforms’ ranking systems may need to distinguish disclosed, useful synthetic media from repetitive AI-generated material; otherwise broad suppression risks penalizing legitimate AI-assisted creators while weak enforcement leaves recommendation quality vulnerable.

Third-order effects

• If fresh-user feeds continue to be flooded with repetitive synthetic video, short-form platforms could shift from optimizing purely for engagement signals toward treating content provenance, repetition, and user controls as ranking inputs.
• The pattern strengthens the case for platform-level disclosure standards and auditable recommendation practices, particularly where children or first-time users are exposed; whether those measures improve feeds depends on enforcement rather than labels alone.