AI Disclosure Labels Drive Down Social Media Engagement

Social media users pull back when AI enters the picture

When content creators disclose they used generative AI to produce their social media posts, audiences respond by disengaging — not because they object to the technology itself, but because they perceive the creator invested less personal effort.

Researchers at the University of Southern California's Marshall School of Business examined how AI disclosure labels affect user behavior on social platforms. The study, conducted by doctoral students Stephan Carney and Ignacio Riveros alongside associate professor Stephanie M. Tully, found that transparency about AI use undermines the perceived authenticity of content.

The mechanism is straightforward: users see an AI disclosure, conclude the creator spent minimal effort on the post, and feel less connection to both the content and its creator. This dynamic holds even though most users understand social media personalities aren't speaking directly to them as individuals.

Why it matters

As generative AI becomes standard in content workflows, this research exposes a tension between transparency and engagement. Platforms and regulators increasingly push for AI labeling, but those disclosures may carry an unintended cost. Brands and influencers who rely on parasocial bonds with audiences — the one-sided emotional connections that drive follower loyalty — face a strategic dilemma: be transparent about AI use and risk lower engagement, or stay silent and potentially face backlash if the truth emerges later.

The authenticity penalty

Social media thrives on parasocial relationships, where audiences develop emotional attachments to creators despite the inherently one-way nature of the connection. These bonds depend heavily on perceived authenticity and effort.

When users learn a post was AI-generated rather than hand-crafted, that perception shifts. The content loses its sense of personal investment, even if the final product is visually or textually identical to what a human might produce. The disclosure itself becomes the problem — not the quality of the output.

Implications for platforms and creators

The findings carry weight for multiple stakeholders. Influencers building personal brands may need to reconsider how and when they deploy AI tools. Brands leveraging AI for scale in their social strategies should account for potential engagement drops. Social media platforms implementing AI labeling policies face questions about whether mandatory disclosure serves user interests if it systematically reduces content interaction.

The research suggests the issue isn't technological skepticism but rather a human preference for content that signals genuine creator investment. As AI generation becomes more sophisticated and harder to detect, the gap between disclosed and undisclosed AI content may widen — creating competitive pressure to avoid transparency.

The study was first reported by Stanford Social Innovation Review and appears in the publication's Fall 2026 issue.