Can you smell what AI is advising? Human, machine, or both

A comparative experimental study on financial advice

Dados Bibliográficos

ID21284095
AutoresSaurav Sathish Kumta (0009-0008-5443-2227, Yenepoya University), Niyaz Panakaje (0000-0003-4568-1658, Yenepoya University, autor correspondente), S M Riha Parvin (0000-0003-4030-6445, Yenepoya University), Ata Mohammad Nikzad (0000-0001-5948-0940), Ata Nikzad (Yenepoya University)
Ano2026
Volume264
Páginas106530
Data de publicação2026-04-01
Peer ReviewedSim
Open AccessSim
TipoARTICLE
PeriódicoActa Psychologica (JOURNAL)
Identificadores do periódicoISSN: 0001-6918 • E-ISSN: 1873-6297
EditoraElsevier BV (PUBLISHER)
DOI10.1016/j.actpsy.2026.106530
PMID41734731
OpenAlexW7131104872
IdiomaEN
Referências citadas65

The study tested if the source of financial advice (Human, AI or Human-AI Hybrid) impacted the outcome of a long-term financial portfolio using a quasi-experimental design with 93 participants. We objectively measured the financial outcomes. By grounding the study in social exchange theory and the Interpersonal Interdependence Framework, we explored the impact of dyadic encounter on trust and acceptance of advice. While every advisor type improved portfolio outcomes, the groups using AI and Hybrid models saw higher success rates compared to the human-only group. While the absolute trust level did not differ across groups, an interesting discovery was how that trust was built. We found that trust in AI isn't an immediate emotional response, but a psychological roadmap where interaction quality builds trust. Participants exhibited less trust towards human advisors compared to AI advisors, as trust is built upon reputation and personal attributes. Our results suggest that AI does not destroy trust, but shifts the reason of trust from personal rapport to process-based interaction. A hybrid model can be a good alternative that can maintain performance of AI while allowing for Human oversight. The future of the financial industry may not have to choose between human or machines. They could potentially utilize AI-based Frameworks by focusing on Calibrated trust

Advice (programming) · Financial services · Interpersonal communication · Interpersonal interaction · Outcome (game theory) · Portfolio · Quality (philosophy) · Reputation · AI in Service Interactions · Decision-Making and Behavioral Economics · Psychology of Moral and Emotional Judgment