Generative AI in Consumer Decision Making: How Algorithmic Recommendations Shape Brand Trust
Year: 2026
Type: article
Abstract: Generative artificial intelligence (GenAI) is rapidly moving from the back end of recommendation engines to the consumer-facing interface, where conversational assistants, generated product descriptions and personalised feeds increasingly decide what shoppers see, compare and buy. While the literature on algorithm aversion and algorithm appreciation explains when consumers accept algorithmic advice, little is known about how such advice reshapes trust in the brand that deploys it. This paper develops and specifies a multi-level, testable framework in which four attributes of a GenAI recommendation (perceived personalisation, perceived accuracy and relevance, transparency and explainability, and disclosure of the AI source) influence brand trust through two opposing appraisal pathways: perceived algorithmic competence and fairness, which build trust, and perceived manipulation and privacy concern, which erode it. Brand trust in turn drives purchase intention and loyalty, with AI literacy and task type (utilitarian versus hedonic) as boundary conditions. Nine hypotheses are derived from trust theory, the persuasion knowledge model and the human–automation literature, and a dynamic trust-calibration loop is proposed in which each decision cycle updates the trust baseline used in the next. A two-study empirical design is specified, comprising a PLS-SEM survey and a 3 × 2 × 2 online experiment, and a Monte Carlo power analysis shows that a survey of about 300 respondents gives at least 95% power for the main hypothesised paths, while a 600-participant experiment gives about 92% power for medium-small effects of disclosure. The paper contributes a brand-level, trust-centred account of GenAI recommendations, identifies transparency as the lever that most plausibly converts algorithmic personalisation into trust, and derives managerial and regulatory implications. The study is conceptual with simulated power analysis; empirical estimates are not reported.
Author: Ms. Manisha Yogesh Thorat
Language: English
Topic: AI in Service Interactions
Subfield: Marketing
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W7225777914
FWCI:
Cites:
Cited by:
Related to: