Consumers More Likely to Lie to Chatbots Than Humans, Study Finds
Quick Look
New research from Sun Yat-sen University suggests consumers are more prone to unethical behavior, like lying or exploiting errors, when interacting with AI chatbots due to reduced social pressure and fear of judgment, a phenomenon termed 'anticipatory face loss'.
AI-generated summary
Why It Matters
New research examines how consumers behave toward artificial intelligence, finding they are more willing to lie to chatbots than customer-service workers. This is attributed to reduced social pressure and fear of judgment when interacting with AI.
People are more willing to lie to a chatbot than a customer-service worker, according to new research examining how consumers behave toward artificial intelligence.
The study, published in the Journal of Business Research, found consumers feel less social pressure or fear of judgment when interacting with AI systems, making unethical behavior more likely.
Researchers from Sun Yat-sen University in China described the effect as “anticipatory face loss,” a term referring to the discomfort people feel when they expect embarrassment or social disapproval.
“Given the high stakes of consumer unethical behavior toward AI and the theoretical gap regarding social face concerns in human-AI interactions, we propose and test the idea that consumers are more likely to engage in unethical behavior toward AI agents than human ones due to reduced anticipatory face loss,” they wrote.
According to the study, consumers were more likely to lie, exploit pricing errors, or falsely claim discounts when interacting with AI instead of humans. In one field experiment, participants also exaggerated results more often for extra rewards when dealing with AI agents.
Researchers found social judgment played a larger role than guilt in unethical behavior toward AI. Participants felt less fear of embarrassment or “losing face” because AI systems were viewed as less socially aware, and less capable of judging them.
The study also found dishonest behavior decreased when AI agents appeared more competent or used eye gaze cues such as simulated eye contact.
The research comes as companies increasingly deploy AI agents for customer support, recommendations, and online transactions. A March 2025 study by research and advisory firm Gartner said AI agents could autonomously resolve 80% of customer-service issues by 2029.
In a separate study by the University of Castilla in October 2025 on humanoid robots, consumers responded more positively to robots with moderate human features, including facial expressions and eye movement, while highly realistic robots often triggered discomfort associated with the “uncanny valley.”
“When robots are anthropomorphized, consumers tend to evaluate the robot more favorably,” the researchers wrote. “Anthropomorphism drives customer trust, intention to use, comfort, and enjoyment. Also, adding human attributes to a robot can make people prefer to spend more time with robots.”
Open Questions
- What specific types of AI agents are most susceptible to unethical behavior?
- Are there cultural differences in how people interact unethically with AI?
- What are the long-term implications of this behavior for businesses?
- How can AI systems be designed to mitigate unethical consumer behavior?
