AI influencer posts gain about 21% more engagement with a real person

By Nathan Brooks

Showing an AI influencer in a genuine social interaction with a real person raised engagement by about 21% in a 2026 study of roughly 10,000 Instagram posts from 28 synthetic creators. The test separated a real exchange from the mere presence of another figure.

What did the study test?

Luca Cascio Rizzo, an assistant professor of marketing, led the research with Francisco Villarroel Ordenes of the University of Bologna and Jonah Berger of the Wharton School.

The team analyzed roughly 10,000 Instagram posts from 28 AI influencers and ran controlled experiments. Their paper, “What drives virtual influencers' impact?”, appeared in the Journal of the Academy of Marketing Science in 2026. It carries the DOI 10.1007/s11747-026-01185-x.

Synthetic creators give brands control: they stay on script, do not age and do not cause scandals. Lil Miquela, one of the best-known examples, has millions of followers and has posted for Prada, Samsung and Dior. Yet audiences do not automatically trust these personalities, which limits the point of looking human.

Cascio Rizzo said: “An AI influencer is, by definition, not human, and audiences know it. But when you show one genuinely interacting with a real person, people start to treat it as more human—and that changes everything.”

Why did the interaction help?

The study's explanation is anthropomorphism, a tendency psychologists describe as people assigning human qualities to nonhuman entities. Posts showing an influencer eating with someone, posing arm in arm or sharing a moment drew more engagement than posts showing the influencer alone.

That perceived human connection can build trust, which can make content more persuasive by encouraging people to pause, engage and remember. The underlying mechanism draws on an old storytelling dynamic: audiences can form attachments to fictional characters.

The study summary draws a parallel with Spike Jonze's film “Her”, where Joaquin Phoenix's character falls for an AI voice assistant through intimate exchanges. The film illustrates the same logic: interaction, not technological novelty, can make an artificial character feel real. It is an illustration, not evidence from the Instagram study.

Does any human presence work?

No. A person standing in the background produced no such effect, and neither did pairing an AI influencer with another AI character. The exchange has to be genuine and meaningful.

Cascio Rizzo said: “It's not about having another face in the frame. It's about signaling a real social connection. That's the cue people read as human.”

In practical terms, a genuine exchange reads like a scene between two people; a background figure reads like scenery. The researcher called authentic human participation a free tactic for brands, and his data show it narrowing the trust gap with human influencers.

Which influencers benefited?

The gain was not uniform. The largest lift came among influencers who looked most artificial. Highly realistic avatars already seemed human enough that a companion added little. The same move did nothing for human influencers because their humanness was never in question.

For marketers, that limits the reach of the tactic. Brands using visibly synthetic characters have more room to improve audience trust than brands whose avatars already pass as human.

What can brands gain?

Only about 15% of consumers say they trust AI influencers, and only about 15% of AI influencer posts show them in a social interaction with a real person. The AI influencer market is projected to surpass $110 billion by 2033. Cascio Rizzo argues that brands are buying control while overlooking a simple trust-building opportunity.

What does the finding not prove?

The study measures whether a synthetic personality seems trustworthy, not whether it deserves trust. Unlike human creators, an AI influencer has never used the products it promotes or accepted consequences for its endorsements. A human bond can make a fabricated person feel credible without proving that its claims are reliable.

This work belongs to a broader research program by Cascio Rizzo and colleagues that uses large-scale data and AI tools to examine how communication and technology shape consumer behavior. As synthetic media moves from novelty into mainstream marketing channels, that research helps explain audience response, but it does not settle whether the appearance of credibility is justified.

The practical gain is measurable; the ethical claim is not. A human companion can improve how an AI influencer lands, not what its endorsement is worth.