AI Workforce Transformation Wiki

Not yet human-reviewed

This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.

Can ChatGPT Be Addictive? A Call to Examine the Shift from Support to Dependence in AI Conversational Large Language Models

Citation. Yankouskaya, A., Liebherr, M., & Ali, R. (2025). Can ChatGPT Be Addictive? A Call to Examine the Shift from Support to Dependence in AI Conversational Large Language Models. Human-Centric Intelligent Systems, 5, 77–89. https://doi.org/10.1007/s44230-025-00090-w

Summary. This peer-reviewed viewpoint paper argues that ChatGPT’s personalization, emotional validation, continuous availability, and support for decision-making may create conditions for behavioral dependence. It frames risks to autonomy, critical thinking, and social relationships as questions for multidisciplinary empirical research rather than demonstrated causal effects. The authors propose adapting behavioral-addiction frameworks and developing measures that can distinguish benefits of use from delayed harms.

Key claims

  • The paper argues that personalized feedback, remembered interaction history, competence support, and immediate responses can reinforce prolonged engagement and potential dependence on ChatGPT.
  • It distinguishes possible ChatGPT dependence from use of narrowly situated AI tools by emphasizing human-like interaction, emotional validation, and broad support for complex decisions.
  • The authors propose research on whether established addiction dimensions such as salience, tolerance, withdrawal, conflict, and relapse capture AI-related dependency and on how personalization changes attachment and decision autonomy.

Evidence & limitations

  • This is a correspondence/viewpoint article that synthesizes prior literature and proposes research questions; it does not estimate the prevalence or causal effects of ChatGPT dependence.
  • The paper identifies methodological challenges including confounding individual differences, simultaneously measuring benefits and harms, and potentially delayed negative effects.

Feeds

  • automation-complacency
  • cognitive-sustainability
  • Proposed new topic: ai-conversational-dependence — behavioral and socio-emotional dependence on personalized conversational AI is central here and is not naturally captured by work-strain or answer-adoption topics.