Evaluating Artificial Intelligence Use and Its Psychological Correlates via Months of Web-Browsing Data

Area of Science:

• Psychology
• Human-Computer Interaction
• Digital Media Studies

Background:

• Public discourse on artificial intelligence (AI) impact is extensive, yet empirical data on real-world AI usage remains scarce.
• Objective measurement of AI adoption in naturalistic settings is needed to understand societal integration.

Purpose of the Study:

• To objectively quantify the frequency of AI use in everyday web browsing among students and the general public.
• To identify psychological and demographic correlates associated with AI adoption.
• To compare self-reported AI use with actual behavioral data.

Main Methods:

• Collected 90-day web browsing data from 499 students (Study 1) and 455 general public participants (Study 2).
• Coded over 4.1 million (Study 1) and 9.9 million (Study 2) websites to identify AI usage.
• Administered psychological trait assessments and demographic questionnaires.

Main Results:

• AI use constituted approximately 1% of student web browsing and 0.44% of general public web browsing.
• Aversive personality traits (Machiavellianism, narcissism, psychopathy) were consistent predictors, though associations varied.
• Demographic factors showed minimal correlation with AI usage.
• Self-reported AI use and actual usage exhibited a moderate correlation (ρ = 0.329).

Conclusions:

• AI adoption in daily web browsing is currently infrequent.
• Individual psychological predispositions, particularly aversive traits, are linked to AI use.
• Behavioral data provides a more accurate measure of AI use than self-reports, highlighting limitations of subjective media use assessments.