Chatbots Offer a Judgment-Free Zone for Sensitive Purchases, Study Finds
Consumers overwhelmingly prefer interacting with clearly non-human chatbots over human representatives when purchasing potentially embarrassing products, according to research led by Jianna Jin, assistant professor of marketing at the University of Notre Dame’s Mendoza College of Business. The findings, published in the Journal of Consumer Psychology, reveal a strong link between a desire to avoid embarrassment adn a preference for machine-like interactions.
jin, along with researchers Jesse Walker and Rebecca Walker Reczek from Ohio State University, conducted seven experiments involving over 6,000 participants to explore how shoppers navigate interactions with chatbots, considering whether the chatbot’s identity is disclosed or ambiguous.The team utilized a custom-built chatbot and presented participants with shopping scenarios involving sensitive items like hemorrhoid medication and anti-wrinkle cream.
One experiment demonstrated that over 80% of consumers seeking diarrhea treatment preferred an online pharmacy utilizing a clearly identified, non-human chatbot compared to one with a human pharmacist. This preference was not observed when shopping for hay fever medication.
Further research, conducted using a dating app scenario, showed a similar trend. Participants faced with sensitive questions about body shape were more likely to disclose data and preferred a clearly non-human chatbot match agent over a human one.
The study also highlighted the impact of ambiguity.When a chatbot’s identity was unclear – appearing and sounding human - consumers prone to embarrassment (like those purchasing hemorrhoid cream) were more likely to assume the agent was human and potentially avoid the interaction altogether, as a protective measure against perceived judgment.
“When it comes to sensitive purchases, if there’s any doubt about who they’re interacting with, people will err on the side of caution and treat the AI chatbot as if it’s human to protect their self-image,” explained jin.”But give them a clearly non-human chatbot, and suddenly that self-presentation pressure vanishes.”
Consumers demonstrated increased willingness to engage with, choose stores utilizing, and even share personal information (like email addresses for free samples) with chatbots they knew were definitively non-human. This suggests a strong preference for machine-like chatbots, especially as the industry moves towards more human-like AI designs.
Jin notes the implications extend beyond typical “self-conscious” purchases, citing potential applications for industries were consumers may face stereotype-based judgment, such as a car leasing company assisting women.The research provides valuable guidance for companies considering chatbot deployment, informing both when to utilize them and how to design them for optimal effectiveness.Source: University of Notre Dame
