Exploring the Factors of AI Chatbot Dependence: An I-PACE Model and AIDUA Framework Perspective
This study utilizes the I-PACE model and AIDUA framework to demonstrate that AI chatbot dependence arises not merely from technology acceptance, but specifically when favorable affective appraisals—driven by factors like hedonic motivation and perceived humanness—transform repeated instrumental use into emotionally reinforced reliance.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the quiet hum of a modern computer, a new kind of conversation is taking place. It is not between two people, but between a human and a machine that speaks with the fluency of a friend. These artificial intelligence chatbots have become common tools for writing, learning, and solving problems. They are designed to be helpful, but their ability to mimic human conversation raises a deeper question: at what point does using a helpful tool turn into an unhealthy reliance? To understand this, researchers look at how our minds process technology. They examine the mix of our personal feelings, our beliefs about how useful a tool is, and the social pressure we feel from others. While we often assume that we use technology simply because it works well, there is a growing concern that the emotional connection we form with these digital voices might be driving us to depend on them in ways that are not always good for us.
A team of researchers from universities in China set out to map the path from casual use to this kind of dependence. They focused on a specific group of people: 540 individuals who had already used AI chatbots. The researchers asked these participants to share their experiences, their feelings, and their habits. They wanted to see if the reasons people started using these bots were the same as the reasons they might become overly attached to them. The study combined two existing ideas about human behavior. One idea suggests that our actions are driven by a cycle of our personal traits, our emotions, our thoughts, and our final choices. The other idea proposes that we judge new technology in steps, first looking at social cues and how much fun it is, then deciding if it is worth the effort, and finally forming an opinion on whether to keep using it. By weaving these two ideas together, the researchers built a model to test what actually pushes a user toward dependence.
The investigation began by looking at the initial reasons people turn to these chatbots. The researchers found that three main factors shape a user's first impression. The first is social influence, or the feeling that friends and colleagues expect you to use the technology. The second is hedonic motivation, which is simply the enjoyment or fun a person gets from using the tool. The third is perceived humanness, the sense that the chatbot understands and responds like a real person. The study revealed that when people find the chatbot enjoyable or feel it is truly human-like, they believe it will help them perform better and require less effort to use. Interestingly, social pressure had a different effect; it made people feel they needed to work harder to use the tool correctly, rather than making the tool seem more useful.
As the users moved from their first impressions to deeper evaluations, the researchers discovered a crucial shift in how the mind works. The participants rated how much they expected the chatbot to improve their work and how much effort it took to use. These practical judgments led to two types of feelings: a logical assessment of the tool's value and an emotional reaction to the experience. The study showed that believing the tool is useful and easy to use leads to both positive logical thoughts and positive feelings. However, when the researchers traced the path all the way to the final outcome—whether a person became dependent on the chatbot or rejected it entirely—they found a surprising result. The logical assessment of the tool's performance did not predict dependence. A person could think a chatbot was highly efficient and still not become reliant on it.
The only factor that directly predicted whether a user would develop a dependence on the chatbot was the emotional reaction. If a user felt a strong positive emotional connection to the interaction, they were far more likely to rely on the chatbot repeatedly, even to the point of maladaptive dependence. This suggests that the transition from using a tool to depending on it is not a matter of cold calculation. It is an emotional process. The study also noted that age and how familiar a person was with the technology played a role, with younger and more familiar users showing different patterns of behavior. Ultimately, the research indicates that while we may start using AI because it is smart or because our friends use it, we stay with it and become dependent on it because of how it makes us feel. The emotional bond, rather than the functional utility, is the engine that drives the shift from a helpful assistant to a necessary companion.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.