Why do AI posts get likes without inviting conversation?

Social proof on social media has historically been a two-stage process. A post is liked or shared (recognition) and is also replied to, quoted, and argued with (engagement). The two stages compound: posts that get replied to tend to be circulated more, and posts that get circulated more tend to get replied to. Influence accrues to authors whose content reliably produces both stages.

AI-generated posts can accumulate the first stage — recognition — at high rates because they are comprehensive, well-formed, and confidently phrased. They cannot easily accumulate the second stage. The post does not invite reply, partly because its register is declarative-without-uncertainty and partly because there is no author present to respond to a counter-claim. So the social proof it earns is one-sided: visibility without conversation.

This produces false social proof in a precise sense. The metric value (likes, shares, saves) implies a kind of community endorsement that the post is not actually receiving, because the community process that would normally validate the metric — argument, response, counter-reply — is suppressed. The numbers compound, but they do not compound on the substrate they were designed to measure.

Two consequences follow. First, recommender systems trained on engagement signals will increasingly optimize for AI-generated content, because the engagement signal it produces is high and cheap. Second, Does AI content displace human influencers on social media? becomes a positive-feedback loop — false social proof crowds out the conversational kind, which produces more false social proof at the expense of the other.

The strongest counterargument: humans have always produced viral comprehensive posts that did not invite reply. True, but at scale that genre was a small share of circulating content and humans paid attention costs to produce it. AI removes the cost and removes the upper limit on share.