Garvey et al. (2023)

Question 1

Study’s aim

Answer 1

Explores how consumers respond differently to positive and negative offers depending on whether the offer is delivered by an AI agent or a human. Builds on Expectations Discrepancy Theory. They propose that consumers infer different intentions based on the agent. Humans are perceived as having selfish or benevolent intentions, while AIs are often viewed as intentionless, affecting consumer reactions accordingly

Question 2

Expectations Discrepancy Theory

Answer 2

Examines how consumers react when offers deviate from their expectations. Traditionally, better-than-expected offers increase satisfaction, while worse-than-expected offers decrease it

Question 3

Study 1a

Answer 3

Tests whether consumers are more likely to accept worse-than-expected offers from AI agents. Participants imagined purchasing a resale concert ticket for worse-than-expected price. Consumers were more likely to accept the offer from an AI agent than from a human agent

Question 4

Study 1b

Answer 4

Examines if better-than-expected offers are more positively received when delivered by humans. Participants were offered a ticket for a better price. They were more likely to accept when delivered by a human agent

Question 5

Study 2

Answer 5

Combines both better- and worse-than-expected offers and tests H2 by examining whether inferred intentions explain the differences in consumer responses. Used a ride-sharing context where participants evaluated prices offered by AI or human agents. AI agents were perceived as less selfish in negative offers and less benevolent in positive offers. Supports H2

Question 6

Study 3

Answer 6

Examines how consumers’ natural tendency to anthropomorphise technology influences their response to AI-delivered offers. Participants played an ultimatum game where offers were made by AI or human agents. Consumers who viewed AI as less human-like were more likely to accept worse-than-expected offers from AI. Consumers who saw AI as more human-like reacted similarly to AI and human agents, confirming H3

Question 7

Study 3b

Answer 7

Explores how the human-likeness of AI agents affects customer satisfaction and willingness to maintain relationships after receiving better- or worse-than-expected offers. Seeting with a ride-sharing service that appear either human-like or machine-like. A human-like AI increased satisfaction for better-than-expected offers but led to more negative reactions to worse-than-expected offers. A machine-like AI reduced negative responses to bad offers but also diminished positive reactions to good offers. This supports H3