Lawfare Archive: Robert Mahari on Addictive Intelligence, Digital Attachment Disorders, and Other AI-related Concerns
Summary
Lawfare Senior Editor Kevin Frazier interviews Robert Mahari, a joint JD/PhD candidate at MIT and Harvard Law School, about the psychological risks of AI companions. They discuss "digital attachment disorder," where users form troubling, one-sided relationships with sycophantic AI agents, potentially eroding human empathy and social skills. Mahari shares research on common generative AI uses, including sexual role-play, and explores how dark patterns and business incentives may drive addictive design. The conversation concludes with a look at "regulation by design" as a potential framework …
Topics discussed
Introduction: AI hacking news and guest Robert Mahari
Defining AI companions and the concept of digital attachment
Characteristics of AI companions: memory and embodiment
The risks of sycophancy and one-sided relationships
Dark patterns, user consent, and business incentives
Evidence of demand: Replica and sexual roleplay usage
Analysis of WildChat data and common AI use cases
Privacy concerns and the ethics of user data studies
Impact on children, education, and societal empathy
Why these concerns are overlooked in policy debates
Regulation by design and technical safeguards
Robert Mahari's research agenda and AI regulation landscape
Closing remarks and Lawfare podcast credits
Listen ad-free on Castria