Is this project an undergraduate, graduate, or faculty project?
Graduate
Project Type
individual
Campus
Daytona Beach
Authors' Class Standing
Brianna Broderick, Senior
Lead Presenter's Name
Brianna Broderick
Lead Presenter's College
DB College of Arts and Sciences
Faculty Mentor Name
Dr. Joseph Keebler
Abstract
Current research on Artificial Intelligence (AI) focuses on its capabilities and our understanding of it as an instrumental tool (i.e., utility completing tasks). However, as its ability to replicate natural language improves through both text and voice, an ever-growing number of users have turned to AI for emotional companionship. Concern grows as prior research on technology dependency suggests AI bonding may lead to less interaction with others and, in extreme circumstances, has already led to cases of suicide and divorce. Kasturiaratna & Hartanto (2025) developed the AI Attachment (AIA) scale, consisting of three factors, which include: emotional closeness (i.e., personal connection and comfort of interacting with AI), social substitution (i.e., relying on the AI for social support/emotional needs), and normative regard (i.e., treating AI with care, politeness, or other morally-infused behavior). This work examined the AIA scale, and, in doing so, identified concerns surrounding item generation, despite the scale’s overall thoroughness. These concerns stem from the scale’s basis in pet attachment theory, which is a divergent construct. In response, this work proposes a nomological network of potential correlations to the AIA scale, including (1) Continuous Usage Intention, (2) Ability to Evaluate AI Language, and (3) Risk of Addiction to AI. While the latter two have not been previously proposed in the literature, this work explores the value of their inclusion. The Ability to Evaluate AI Language construct corresponds to users’ evaluation of emotional meaning rather than analytics or comparison between AI models, as currently emphasized in AI Literacy. Similarly, the Risk of Addiction to AI construct is based on the current understanding of impulse control disorders and scales of technological addiction. Future steps for this project involve developing a novel scale centered on the Risk of Addiction to AI construct to support the necessary research on socioemotional usage of AI.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Assessing Attachment to AI: Understanding the Theoretical Correlations and Consequences
Current research on Artificial Intelligence (AI) focuses on its capabilities and our understanding of it as an instrumental tool (i.e., utility completing tasks). However, as its ability to replicate natural language improves through both text and voice, an ever-growing number of users have turned to AI for emotional companionship. Concern grows as prior research on technology dependency suggests AI bonding may lead to less interaction with others and, in extreme circumstances, has already led to cases of suicide and divorce. Kasturiaratna & Hartanto (2025) developed the AI Attachment (AIA) scale, consisting of three factors, which include: emotional closeness (i.e., personal connection and comfort of interacting with AI), social substitution (i.e., relying on the AI for social support/emotional needs), and normative regard (i.e., treating AI with care, politeness, or other morally-infused behavior). This work examined the AIA scale, and, in doing so, identified concerns surrounding item generation, despite the scale’s overall thoroughness. These concerns stem from the scale’s basis in pet attachment theory, which is a divergent construct. In response, this work proposes a nomological network of potential correlations to the AIA scale, including (1) Continuous Usage Intention, (2) Ability to Evaluate AI Language, and (3) Risk of Addiction to AI. While the latter two have not been previously proposed in the literature, this work explores the value of their inclusion. The Ability to Evaluate AI Language construct corresponds to users’ evaluation of emotional meaning rather than analytics or comparison between AI models, as currently emphasized in AI Literacy. Similarly, the Risk of Addiction to AI construct is based on the current understanding of impulse control disorders and scales of technological addiction. Future steps for this project involve developing a novel scale centered on the Risk of Addiction to AI construct to support the necessary research on socioemotional usage of AI.