Claim
Platform detection systems now restrict AI-generated content's algorithmic reach before removal, making distribution access, not production cost, the binding constraint for content at scale.
Mechanism
Platforms deploy detection layers (LinkedIn, YouTube, Reddit, TikTok, Substack) that identify AI-generated or low-effort content and cap its reach to the creator's immediate network rather than deleting it. Production volume is no longer the binding resource. The production cost advantage AI provides is absorbed by distribution suppression before content reaches the broader algorithmic feed. Scarcity migrates: "can I create enough?" becomes "can I reach anyone?" Quality signals, original perspective, and primary data determine algorithmic distribution; volume alone does not.
Conditions
Holds when: creating content for platforms with active detection (LinkedIn, YouTube, Reddit, TikTok, Substack). Applies to content that is formulaic, repetitive, or lacks original data or perspective.
Fails when: content carries a genuine original perspective, a personal voice, or primary data that passes detection. Fails for channels without algorithmic gatekeeping, such as email newsletters and direct messaging.
Evidence
Pangram Labs scanned more than one million LinkedIn posts from April through June 2026 and found 41% of long-form posts and 23% of comments were fully AI-generated. LinkedIn detection accuracy is approximately 94%, implying roughly one in 17 flagged items is a false positive. LinkedIn is testing a user-reporting button labeled "Seems like AI Slop" and replaced its "Enhance Post" feature with grammar-only proofreading. Posts exceeding detection thresholds are capped to the creator's immediate network. YouTube demonetizes "repetitive, low-effort" content. Reddit deploys AI-based spam detection. TikTok requires AI labels with embedded metadata watermarks. Substack uses Pangram detection on its platform.
Indig's conclusion is that production efficiency gains from AI do not offset losses in distribution reach when platforms apply active suppression.
Signals
- AI-assisted posts receive engagement from first-degree connections but no amplification beyond them.
- Algorithmic reach of AI-heavy content falls materially compared to posts written in a personal voice.
- Content appears on your profile page but not in broader algorithmic feeds.
Cross-references
- A single seeded fake claim can self-confirm in AI Overviews: Lily Ray's case shows AI-generated content can self-propagate through AI Overviews, the inverse of Indig's distribution-suppression finding. The two mechanisms can compound: platform suppression limits reach while AI Overview propagation locks in false claims.
- UGC platforms hold 4x more AI-cited domains than publishers; community signals are AI search's largest unoptimized variable: UGC platforms that Indig shows dominate AI citations are now deploying detection against AI-generated UGC, compressing the citation advantage from both ends simultaneously.