aceModerate evidence
More
AI as persuasive as humans overall, but weaker on intentions, review of 121 experiments finds
Study title: Is artificial intelligence more persuasive than humans? A meta-analysis
- Evidence type
- Systematic review or meta-analysis
- Sample
- n = 53,977
- Sources
- 1 sourcePeer-reviewed
- Published
- 1 min read
- Truvace record
- TRV-2026-2735cite or check it
Study title: Is artificial intelligence more persuasive than humans? A meta-analysis
A meta-analysis in the Journal of Communication pooled 121 randomized experiments with 53,977 participants that compared AI and human communicators on persuasion outcomes. Overall, AI agents were as persuasive as humans.
Read the study abstract
Abstract The rapid deployment of artificial intelligence (AI) technology has enabled AI agents to take on various roles as communicators, such as virtual assistants, robot journalists, and AI doctors. This study meta-analyzed 121 randomized experimental studies (N = 53,977) that compared the effects of AI and human agency on persuasion outcomes, including perceptions, attitudes, intentions, and behaviors. The results showed that AI agents were as persuasive as humans in terms of overall persuasion outcomes. With…
The problem
A meta-analysis of 121 randomized experiments (N = 53,977) found AI agents were less effective than humans at shaping people's behavioral intentions.
In brief
Results varied by outcome type. AI was less effective than humans at shaping behavioral intentions, but did not differ significantly from humans in changing perceptions, attitudes or actual behaviors.
The authors also found uneven patterns across AI communicator roles, directions of communication, experimental settings and demographic groups, so the overall parity does not hold in every context.
Main points
- The analysis combined 121 randomized experimental studies with 53,977 participants in total.
- Studies compared AI and human agency on perceptions, attitudes, intentions and behaviors.
- Overall persuasion outcomes did not differ between AI agents and humans.
- AI was less effective than humans at shaping behavioral intentions.
- No significant AI-human difference appeared for perceptions, attitudes or actual behaviors.
- Effects differed by AI communicator role, direction of communication, experimental setting and demographic segment.
The full story
The study examined AI agents acting as communicators, including virtual assistants, robot journalists and AI doctors, and asked whether they persuade as well as humans do. It pooled randomized experiments that directly compared AI with human agency.
The pooled result was parity on overall persuasion. Breaking results down by outcome, only behavioral intentions showed a gap, with AI performing worse, while perceptions, attitudes and actual behaviors showed no significant difference.
The authors report heterogeneous patterns by AI role, direction of communication, experimental setting and demographic segment, and discuss implications for human-machine communication and persuasion.
Sources
1 on file- Journal of Communicationdoi:10.1093/joc/jqad024
Talk it over
Who carries this cost, and what would have to change to prevent it?
AI as persuasive as humans overall, but weaker on intentions, review of 121 experiments findshttps://truvace.com/item/is-artificial-intelligence-more-persuasive-than-humans-a-meta-analysis
Who carries this cost, and what would have to change to prevent it?
The problem A meta-analysis of 121 randomized experiments (N = 53,977) found AI agents were less effective than humans at shaping people's behavioral intentions.
Answer in the threadHow do we know? What can this kind of evidence show, and what can’t it?
The evidence This rests on a systematic review or meta-analysis with a sample of 53,977, from Journal of Communication.
Answer in the threadWhat does it leave open? What else would you want to know before acting on it?
The record’s own caveat The authors report heterogeneous patterns across AI roles, communication directions, experimental settings and demographic segments, so the pooled results do not apply uniformly.
Answer in the threadHave you seen AI used this way where you study, work or live? Did it go the same way?
Answer in the thread
The questions are the same on every record. The details under them are taken from this record as written.
What readers say