Review of 45 studies finds AI's downsides hurt customer trust and sales

The harms reached customers' thinking (perceived benefit, trust), feelings (attitude, satisfaction) and behaviour (purchase, loyalty, well-being). The authors describe the work as among the first to synthesise research on the dark side of AI in marketing.

The size of the effects varied with context: whether AI was online or offline, customer age and gender, product type, firm type and national culture all moderated the relationships. The authors say marketing managers should weigh these drawbacks and policymakers should consider vulnerable groups.

The researchers synthesised effect sizes from earlier studies rather than running a new experiment. The pool covered 45 studies and 50 independent samples with 19,503 participants in total, aimed at the negative side of AI that they say has had limited attention compared with its benefits.

Adverse effects were grouped as privacy concern, perceived risks, customer alienation and uniqueness neglect. These lowered customers' cognitive responses (perceived benefit, trust), affective responses (attitude, satisfaction) and behavioural responses (purchase, loyalty, well-being).

Moderators shaped the strength of these effects, including AI setting (online or offline), customer age and gender, product type, level of involvement, firm sector and national cultural dimensions. The authors recommend that managers find ways to minimise these effects in different contexts and that policymakers pay attention to vulnerable groups.