As smart devices become more advanced, companies increasingly design them to exhibit distinct personalities. A smart speaker might act as an obedient assistant, a learning app might behave like a collaborative peer, and a chess robot might present itself as a formidable rival. Researchers wanted to know how these different social roles influence a consumer’s willingness to purchase a product.
A recent study published in Psychology & Marketing investigated this dynamic by testing how different artificial intelligence personas affect human buyers. The research team explored whether framing an AI as a servant, a partner, or a competitor shifts how people perceive their own standing relative to the machine. They found that a consumer’s decision to buy an AI product is heavily influenced by how much the device threatens their sense of dominance.
The Human Need for Dominance
The investigation was led by Crystal T. Lee of Shantou University and Yung-Cheng Shen of National Taiwan Normal University. The researchers based their work on social dominance theory, which suggests that people naturally prefer to maintain clear social hierarchies. In the context of technology, this means humans generally want to remain in a dominant position over the tools they use.
The researchers theorized that AI social roles act as signals about the intended power dynamic between the machine and the user. A servant AI signals submission, a partner AI signals equality, and a competitor AI signals a direct challenge to human capabilities. The team designed four separate experiments to see how these signals play out in simulated retail environments.
In the first experiment, 634 participants evaluated a fitness robot in a simulated online shopping environment. The robot was randomly presented to different groups as either a servant, a partner, or a competitor. The researchers then asked participants to rate their feelings of threat, their sense of dominance, and their likelihood of buying the product.
The results revealed a specific sequence of psychological reactions. When the AI was positioned as a competitor, participants reported the highest levels of threat. This heightened threat perception diminished their sense of dominance, which ultimately resulted in the lowest purchase intentions among the three groups.
Conversely, the servant AI generated the lowest feelings of threat. This allowed participants to feel a strong sense of dominance over the technology. As a result, the servant AI generated the highest purchase intentions, followed closely by the partner AI.
Softening the Threat With Design
The researchers next wanted to see if they could alter these psychological reactions by changing the physical appearance of the product. In a second experiment, 637 participants evaluated an AI programming robot. The researchers manipulated both the social role of the AI and its visual design, presenting the robot with either sharp, angular features or soft, curved lines.
Prior psychological research indicates that angular shapes often trigger subconscious threat detection, while rounded shapes signal safety. The experiment showed that a curved design significantly softened the threat level of a competitor AI. When the challenging AI was encased in a rounded, friendly looking exterior, participants were much less defensive than when it featured sharp angles.
For the servant AI, the visual design made very little difference. Because the servant role already minimized feelings of threat, changing the robot’s physical shape provided diminishing returns. The visual design acted primarily as a buffer for the more aggressive AI roles.
The Problem With Sarcastic Servants
A third experiment tested how communication styles interact with these social roles. The researchers had 643 participants interact with the AI products, this time manipulating the type of humor the AI used. The AI deployed either affiliative humor, which is friendly and inclusive, or aggressive humor, which relies on sarcasm and teasing.
The researchers found that aggressive humor severely damaged the appeal of the servant AI. Consumers expect service-oriented agents to exhibit deference and subordination. When a servant AI used sarcastic or teasing humor, it violated those expectations and caused a sharp decline in the user’s sense of dominance.
In contrast, the partner AI was largely immune to the negative effects of aggressive humor. Because the partner role establishes an equal relationship, participants interpreted the aggressive humor as playful banter between peers. The threat of a sarcastic joke was perceived as benign when it came from a collaborative partner.
When Consumers Want a Challenge
While the first three experiments suggested that competitor AI generally performs poorly in the broad consumer market, the researchers identified one major exception. In a fourth experiment, 627 participants evaluated an AI board game robot. This time, the researchers measured the participants’ mastery goals, a psychological trait reflecting an individual’s drive to develop new skills and conquer difficult challenges.
The results completely reversed the previous patterns for a specific group of people. Participants with high mastery goals actually preferred the competitor AI over the servant and partner models. Instead of avoiding the high-threat machine, these individuals interpreted the challenge as a valuable opportunity for self-improvement.
Participants with low mastery goals behaved exactly as the broader population had in earlier tests. They avoided the competitor AI and showed a strong preference for the supportive, non-threatening servant AI. The researchers interpret this as evidence that competitive AI positioning can be highly effective, but only when targeted at the right audience.
Strategies for AI Adoption
These findings offer concrete strategies for businesses developing and marketing AI products. For companies pursuing broad market adoption, the default positioning should emphasize service and support. Products designed for the general public are most likely to succeed when they explicitly reinforce the user’s sense of control.
However, companies targeting high-performance niches can benefit from a competitive positioning strategy. For products aimed at professionals in competitive fields or individuals seeking advanced skill development, marketing an AI as a formidable challenger can increase user engagement. The research indicates that this positioning should be treated as a segmentation strategy rather than a universal approach.
Firms can also use design elements to balance their product messaging. If a product must function as a competitor, designers can use curved, softer physical features to mitigate the user’s defensive responses. Additionally, companies should ensure that AI assistants programmed to act as servants stick strictly to warm, affiliative communication styles.
This dynamic extends beyond consumer products to the integration of AI in the workplace. Organizations introducing new AI tools might assess the goal orientations of their workforce before rolling out the technology. For employees focused on skill acquisition, introducing the AI as a challenging counterpart can drive engagement. For employees with lower mastery orientation, positioning the AI as a supportive assistant can help minimize resistance across a diverse workforce.
