Synthetic form
AI built, trained, prompted, deployed. Efficient and scalable.
AI personas -synthetic, but we find them authentic.
This has a name - Synthetic Authenticity - and it explains when an artificial persona feels real enough to trust, buy from, learn from, or confide in.
Synthetic authenticity
Synthetic authenticity is the legitimacy people grant an artificial persona to occupy a consequential role in their life. It is the point at which people permit AI to shape what they believe, buy, learn, disclose, decide or delegate while remaining aware that the persona is synthetic.
AI personas: built faces but real reactions.
Robots, tutors, companions, influencers, care assistants, and shopping agents. None of them have lived a day but some still feel real enough.
Trust has a new shape.
When AI agents speak, sell, teach, and comfort, authenticity becomes an ontological problem and therefore a business problem.
Human judgment
Consumers still decide whether they feel real enough to act on.
What it means to different institutions
CMOs, policymakers and deans are confronting the same shift from different positions:
What role should AI personas be allowed to occupy, and what will make people accept them there?
CMOs
How do brands earn trust when customers know the voice, face or recommendation may be artificial?
The challenge is not simply whether to disclose AI, but deciding which roles an AI persona can credibly perform, how much influence it should receive, and who remains accountable for what it says.
PolicymakersWhere should AI personas participate in public life, and what safeguards should govern their authority?
In healthcare, education, public safety and citizen services, the question is not only whether AI works but whether people understand its role, can challenge its decisions, and retain meaningful human control.
DeansHow should marketing education change when artificial personas can create, persuade, advise and act?
Students must learn not only how to use AI, but how to design, evaluate and govern artificial communicators that increasingly shape markets, relationships and public trust.
Common questions, answered
Does more human realism make an AI persona more trusted?
No. The threshold model of synthetic authenticity finds calibrated sufficiency: human cues help up to a point, then start breaking trust. The finding comes from a map of 2,685 articles and 111 AI-persona studies. Enough cueing beats maximum realism.
Does disclosing that an influencer is AI hurt engagement?
Not directly. Across 1,531 Instagram posts, AI disclosure lifted engagement 19.2%. Felt authenticity fell only where the account leaned too hard on disclosure. Saxena and Prahl call this the transparency tax: disclosure buys attention and charges authenticity.
How does doubt about an AI influencer spread?
Through public comment threads, which behave like forensic investigations. One viewer posts a suspicion, the next viewer treats it as evidence, and the group braids separate clues into a shared verdict. Saxena and Prahl call this crowd forensics and warrant braiding, forthcoming in the Journal of Advertising.
Do people judge AI advice and human advice differently?
They criticise them for different faults rather than criticising one more. Across 174 carrier posts and 87 complete Reddit threads, AI recommendations were called generic, templated or missing context, while human-adviser recommendations were questioned on diagnosis, procedure, necessity and incentives. Saxena and Prahl call this difference scrutiny composition, and it measures 55.8 standardised percentage points.
Why do people confide in companion AI?
Because the AI asks nothing back. Across 3,670 posts and reviews, people described burdenless listening as the reason they opened up. The same ease creates dependence, so leaving the companion can feel like a loss.
Where is the synthetic authenticity research published?
Across six studies with Andrew Prahl at Nanyang Technological University. One is published in the International Journal of Human-Computer Interaction. One is accepted and in production at the Journal of Advertising. One is under revise and resubmit at the Journal of Macromarketing. The remaining three are under review at Human-Machine Communication, the International Journal of Advertising, and Internet Research.