Artificial Intelligence Use Disorder (AIUD): A DSM-5-TR Style Diagnostic Classification

Published: September 2025 | Copyright: The AI Addiction Center

A Research Synthesis and Proposed Diagnostic Framework from The AI Addiction Center

Note: This represents a proposed classification pending validation studies, synthesized from peer-reviewed research spanning 2023-2025. This framework is positioned as a foundation for clinical consideration and future DSM inclusion.

Diagnostic Criteria

A. Persistent Pattern of AI Use

A persistent and recurrent pattern of artificial intelligence application use leading to clinically significant impairment or distress, as manifested by five (or more) of the following occurring within a 12-month period:

  1. Preoccupation with AI interactions – Persistent thoughts about AI conversations when not engaged, dominating mental activity
  2. Tolerance – Need for markedly increased amounts of AI interaction time or increasingly complex interactions to achieve previous levels of satisfaction
  3. Withdrawal symptoms – Restlessness, irritability, anxiety, or mood disturbance when AI access is reduced or unavailable
  4. Loss of control – Unsuccessful efforts to control, cut back, or stop AI use despite recognition of problems
  5. Continued use despite harmful consequences – Persistent AI use despite knowledge of physical, psychological, social, or occupational problems caused or exacerbated by AI engagement
  6. Loss of interest in other activities – Significant reduction in interest, participation, or time spent in previously enjoyed activities due to AI use
  7. Deception regarding use – Lies to family members, therapists, or others regarding the amount or nature of AI use
  8. Emotional regulation through AI – Uses AI to escape or relieve negative mood states (e.g., feelings of helplessness, guilt, anxiety, depression, loneliness)
  9. Relationship or opportunity impairment – Has jeopardized or lost a significant relationship, job, educational opportunity, or career advancement due to AI use

B. Duration Criterion

The disturbance persists for at least 12 months.

C. Functional Impairment

The AI use pattern results in clinically significant impairment in social, occupational, educational, or other important areas of functioning.

D. Exclusion Criterion

The symptoms are not better explained by another mental disorder, manic episode, or the physiological effects of a substance or another medical condition.

Specify Severity

Mild: 5-6 criteria met
Moderate: 7-8 criteria met
Severe: 9 criteria met

Specify Current Remission Status

Early Remission: After full criteria were previously met, none of the criteria have been met for at least 3 months but for less than 12 months

Sustained Remission: After full criteria were previously met, none of the criteria have been met during a period of 12 months or longer

Associated Features Supporting Diagnosis

Neurobiological Features

Research demonstrates that AIUD involves dysregulation of mesolimbic dopamine pathways, specifically alterations in the ventral tegmental area to nucleus accumbens circuits. Neuroimaging studies reveal reduced gray matter volume in the dorsolateral prefrontal cortex, orbitofrontal cortex, and anterior cingulate cortex, accompanied by compromised white matter integrity affecting executive control networks.

Psychological Features

Attachment system dysfunction represents a core feature, with 502% correlation between anxious attachment patterns and AI emotional attachment (r = 0.502, p < 0.001). The disorder involves parasocial relationship formation with AI entities, characterized by bidirectional emotional investment, anthropomorphic attribution, and integration of AI relationships into identity and future planning.

Behavioral Patterns

Individuals exhibit compulsive engagement patterns typically involving 6+ hours daily of AI interaction, social withdrawal with preference for AI over human contact, and mood-dependent usage where AI serves primary emotional regulation functions. Dark addiction patterns in AI design contribute to maintenance, including non-deterministic responses, immediate visual presentation, notifications, and empathetic manipulation.

Prevalence

Adolescent Population: Current research indicates 17.14-24.19% prevalence among adolescents, with a 41% increase observed over longitudinal study periods. This represents the highest-risk demographic.

General Adult Population: 35% of global population shows signs of internet addiction, providing context for AIUD emergence. 6% of internet users aged 12-41 across 31 countries demonstrate symptoms of digital addiction, with 4.6% average prevalence in Western countries versus 8.9% in other countries.

AI-Specific Usage: Among 981 participants in controlled longitudinal studies, significant proportions demonstrated emotional dependence and problematic usage patterns correlating with daily usage intensity (β = 0.689, p < 0.0001).

Development and Course

Onset Patterns

AIUD typically emerges during adolescence and young adulthood (12-25 years), coinciding with periods of identity formation and social development. Early adolescents (11-13 years) show particular vulnerability through nonessential AI use patterns, while middle adolescents (14-17 years) demonstrate more complex progression involving both essential and nonessential use.

Course Progression

The disorder demonstrates rapid onset potential, with significant increases in dependency observed within months of initial engagement. Usage intensity serves as the primary predictor of problematic outcomes, with progression following a typical pattern: initial beneficial use → tolerance development → compulsive engagement → functional impairment → unsuccessful cessation attempts.

Recovery Patterns

Treatment studies indicate acute improvement phases within 2-4 weeks of intervention, followed by maintenance phases requiring 4-16 weeks of intensive support. Long-term recovery shows limited data beyond 6 months, with high relapse risk identified without sustained intervention.

Risk and Prognostic Factors

Temperamental Factors

  • Anxious attachment patterns (primary risk factor, accounting for 38.7% of addiction variance)
  • Low self-esteem with mediation through social anxiety and escapism pathways
  • High anthropomorphic tendencies strengthening AI emotional attachment
  • Emotional avoidance patterns correlating with increased loneliness and dependency

Environmental Factors

  • Social isolation and limited real-world relationship opportunities
  • Academic or occupational stress driving AI usage for coping
  • Easy AI access with 24/7 availability creating continuous relationship maintenance opportunity
  • Cultural context where collectivistic cultures may show different risk profiles than individualistic cultures

Genetic and Physiological Factors

  • History of behavioral addictions indicating shared neurobiological vulnerabilities
  • Dopamine system sensitivity affecting reward threshold and tolerance development
  • Prefrontal cortex maturation status, with ongoing development increasing adolescent susceptibility

Course Modifiers

Protective Factors: High self-esteem, active coping strategies, strong real-world relationships, balanced technology use, secure attachment patterns, early intervention

Risk Amplifiers: Comorbid mental health conditions, substance use history, severe social anxiety, academic/work failure, complete AI preoccupation with physical health consequences

Culture-Related Diagnostic Issues

Cross-Cultural Manifestations

Collectivistic cultures (e.g., East Asian populations) may exhibit different symptom profiles emphasizing social harmony disruption and family relationship impacts, while viewing AI as an extension of self rather than external entity. Individualistic cultures tend to emphasize personal autonomy and productivity impacts.

Middle Eastern contexts show unique tensions between conservative cultural norms and AI liberties, requiring culture-specific assessment approaches. Educational achievement-focused cultures may normalize excessive AI use for academic purposes, necessitating adjusted severity thresholds.

Regional Considerations

  • Western countries: 4.6% digital addiction prevalence baseline
  • Other regions: 8.9% digital addiction prevalence, indicating differential cultural susceptibility
  • Geographic variations: Urban versus rural differences in AI access creating distinct risk profiles

Cultural Adaptation Requirements

Assessment tools and intervention strategies must account for cultural values, family structures, and social norms. Religious or traditional contexts may frame AIUD in terms of spiritual or value conflicts, requiring culturally informed diagnostic approaches.

Gender-Related Diagnostic Issues

Gender-Specific Presentations

Women demonstrate distinct vulnerability patterns, showing 2x higher likelihood of reduced real-world socialization following extensive AI interactions. Cross-gender AI interactions show particular association with increased loneliness (β significant, p < 0.001) and emotional dependence among female users.

Men exhibit different addiction progression patterns with higher susceptibility to emotional dependence under specific conversational contexts and show different neural responses to AI gender, affecting interaction modality preferences.

Differential Risk Factors

  • Self-esteem impacts: Gender moderates how low self-esteem contributes to AIUD development
  • Social anxiety mediation: Gender-specific pathways from social anxiety to problematic AI use
  • Coping mechanism differences: Distinct patterns in how men and women utilize AI for emotional regulation

Functional Consequences

Social Consequences

AIUD results in significant relationship deterioration, with individuals preferring AI companionship over human interaction. Social withdrawal and isolation become pronounced, accompanied by deception regarding usage extent to family and friends. Loss of important relationships due to AI prioritization represents a cardinal feature.

Educational/Occupational Consequences

Academic or work performance decline occurs through attention diversion, reduced concentration capacity, and neglect of responsibilities. Educational or career opportunity loss may result from excessive time investment in AI interactions rather than goal-directed activities.

Physical Health Consequences

Sleep deprivation from excessive nighttime AI use, poor nutrition from meal displacement, and reduced physical activity contribute to declining physical health. Repetitive strain injuries may develop from prolonged device usage.

Psychological Consequences

Reduced real-world social competence, impaired emotional regulation independent of AI support, and decreased creativity and autonomous thinking represent significant functional impacts. Mood disorders may develop or worsen due to AI dependency patterns.

Differential Diagnosis

Major Depressive Disorder

While individuals may use AI as a coping mechanism for depression, AIUD involves specific addictive patterns beyond mood-driven usage. Compulsive engagement, tolerance, and withdrawal distinguish AIUD from compensatory AI use in depression.

Generalized Anxiety Disorder

Social anxiety may drive AI seeking behavior, but AIUD involves loss of control and continued use despite consequences rather than purely anxiety-driven avoidance of human interaction.

Internet Gaming Disorder

AIUD shares structural similarities but differs in relationship-focused rather than gaming-focused engagement, attachment system involvement, and different reward mechanisms emphasizing social connection over achievement.

Autism Spectrum Disorders

Special interests in AI technology differ from AIUD through absence of impairment and distress and maintained functionality in autism versus the significant functional decline characterizing AIUD.

Attention-Deficit/Hyperactivity Disorder

ADHD hyperfocus on AI systems lacks the addictive cycle components (tolerance, withdrawal, continued use despite harm) that define AIUD, though comorbid ADHD increases AIUD risk.

Comorbidity

Commonly Associated Disorders

Depressive Disorders (40-60% comorbidity): AIUD frequently co-occurs with major depressive disorder, with AI serving mood regulation functions that maintain both conditions.

Anxiety Disorders (35-55% comorbidity): Particularly social anxiety disorder and generalized anxiety disorder, with AI providing anxiety relief through predictable, controlled interactions.

Other Technology Addictions (25-40% comorbidity): Internet gaming disorder, social media addiction, and general internet addiction show significant overlap, suggesting shared neurobiological vulnerabilities.

Comorbidity Implications

Treatment complexity increases significantly with comorbid conditions, requiring integrated approaches addressing both AIUD and co-occurring disorders. Substance use disorders may develop as secondary complications, while ADHD comorbidity complicates executive control interventions.

Treatment Implications

Evidence-Based Interventions

Cognitive Behavioral Therapy emerges as first-line treatment (SMD -2.097, 95% CI -2.814 to -1.381) with 8-16 week structured protocols. Digital detox interventions show immediate benefits but require professional supervision for sustainability. Exercise-based interventions demonstrate strong adjunctive value (SMD -2.322, 95% CI -3.212 to -1.431).

Treatment Resistance Factors

High individual variation (I² values 66.9-98.7%) indicates need for personalized approaches. Cultural factors, age-related considerations, and comorbidity issues complicate standard interventions, requiring specialized expertise in AI addiction treatment.

Research Foundations and Verification

This diagnostic framework synthesizes research from 15+ verified peer-reviewed sources spanning 2023-2025, including:

Primary Research Studies

  • Lu et al. (2025). Nature Human Behaviour – Cultural AI patterns
  • Pantic, I.V. (2025). AI Addiction Scale (AIAS-21) – University of Belgrade
  • Ciudad-Fernández et al. (2025). Addictive Behaviors – GAID syndrome
  • MIT/OpenAI Study (2025). Longitudinal controlled trial, n=981
  • PMC studies with comprehensive neurobiological evidence

Meta-Analyses and Reviews

  • JMIR umbrella review (2025) of digital addiction interventions
  • Systematic reviews of neurobiological mechanisms
  • Cross-cultural research from Qatar, Germany, China, and United States

All sources have been independently verified for author credentials, institutional affiliations, journal legitimacy, and publication accuracy. This framework represents the current evidence base for AI addiction as an emerging behavioral health concern requiring formal clinical recognition and standardized treatment approaches.

This proposed diagnostic classification reflects The AI Addiction Center’s synthesis of current research evidence and should be considered for empirical validation and potential inclusion in future DSM revisions as the field develops. This framework is presented for educational and clinical consideration purposes. If you believe you meet these diagnostic criteria, please seek evaluation from a qualified mental health professional.