As AI chatbot use has exploded, peer-reviewed studies examining their addictive potential have begun appearing in scientific journals. These published findings — vetted through peer review — contribute to the growing evidence base around AI chatbot use patterns. Several findings have emerged that researchers are examining further.

Key Published Findings

Peer-reviewed studies have explored several patterns, though findings remain preliminary:

  • Researchers found that users who engage with AI chatbots for emotional support tended to show higher rates of problematic use than those using AI for utilitarian purposes
  • Some studies suggested that younger users, particularly adolescents and young adults, may show higher vulnerability to chatbot dependency
  • Pre-existing loneliness and social anxiety appeared to be associated with AI chatbot dependency in several studies
  • The perceived empathy of AI chatbots appeared to correlate with attachment strength and dependency risk
  • Researchers observed that users often underestimated their own level of AI chatbot engagement when self-reporting

Methodological Approaches

Published studies employ various methodologies: cross-sectional surveys measuring AI use patterns and psychological correlates, experimental designs testing the effects of AI interaction on mood and behavior, qualitative studies exploring user experiences, and computational analyses of usage data.

What the Evidence Supports

The peer-reviewed evidence currently suggests that AI chatbot use may become problematic for a subset of users, that the mechanisms may share features with other compulsive behavior patterns researchers have studied, and that certain design features could increase dependency risk. The evidence base is growing but remains limited by the novelty of the technology.

Gaps in the Literature

Significant gaps remain: few longitudinal studies, limited intervention research, under-representation of diverse populations, and insufficient focus on specific AI platforms and their unique risk profiles. Addressing these gaps is a priority for the research community.

Want to learn more about AI dependency? Visit AI Am Addicted for awareness resources and self-reflection tools.