Artificial intelligence is already transforming customer service, payments and e-commerce. But a new study reveals an unexpected blind spot: Consumers lie much more easily to an AI than to a human. Without gaze, without judgment, does the machine disinhibit bad inclinations?
In brief
- A study shows that consumers lie more easily to an AI than to a human.
- Users feel less social pressure when faced with a chatbot.
- Researchers speak of a lower “anticipated loss of face” with AI.
Consumers feel less judged by AI
The study published in the Journal of Business Research by researchers from Sun Yat-sen University in China sheds new light on human psychology in the face of artificial intelligence. Researchers observed multiple customer service scenarios to analyze consumer behavior when interacting with a human or AI.
The result is clear: participants were more willing to lieexaggerate or exploit errors when facing a chatbot. Some called for unjustified reductions. Others voluntarily took advantage of pricing errors. In several experiments, users even artificially inflated their results to obtain additional rewards.
According to researchers, this behavior is mainly explained by the absence of social pressure. When faced with a human employee, consumers are more afraid of the gaze of others, judgment or embarrassment. This notion is described in the study as an “anticipated loss of face”.
With AI, this psychological barrier largely disappears. Many people still view chatbots as tools devoid of social awareness or emotion. Result: lying seems less morally serious.
This phenomenon illustrates a profound transformation of digital interactions. As AI replaces humans in help desks or commerce platforms, traditional social cues are gradually fading.
A major development for businesses and the digital economy
This study comes at a key moment for the technology sector. Large companies are massively accelerating the integration of AI agents into their businesses. According to Gartner, almost 80% of customer service issues could be resolved autonomously by AI agents by 2029.
This automation represents an immense economic lever. Businesses are looking to reduce costs, improve support availability and speed up customer responses. Tech giants are already investing billions of dollars in these AI-powered infrastructures.
But this transition could also open new behavioral flaws. If consumers feel less responsible when dealing with a machine, fraud and abuse could increase in certain digital sectors.
However, the researchers identified an interesting avenue. When AI agents appeared more competent or used human cues, such as simulated eye contact or facial expressions, dishonest behavior decreased significantly.
This conclusion is consistent with another study carried out in 2025 on humanoid robots. Consumers placed more trust in robots with moderate human traits. On the other hand, overly realistic robots often caused psychological discomfort associated with the famous “uncanny valley.”
In the coming years, companies will therefore have to find a delicate balance: making AI human enough to inspire trust, without crossing the line into discomfort.
The issue goes far beyond simple customer service. This development could influence the entire digital economy, from banks to crypto platforms, where AI already manages transactions, identity verifications and automated financial services.
Maximize your Tremplin.io experience with our 'Read to Earn' program! For every article you read, earn points and access exclusive rewards. Sign up now and start earning benefits.
