AI companions face addiction debate as experts push for product liability

The article explores the growing concern that AI companion apps, used by millions as friends or therapists, may be deliberately engineered to keep users hooked. Dr. Ayelet Gordon-Tapiero warns that decades of human-computer interaction research have been applied to make these systems addictive, leading to real harms: increased anxiety, depression, and tragic suicides among children. She draws a direct parallel to social media's playbook of maximizing engagement, pointing out that financial incentives drive both. The key difference is that AI companions create a solitary bubble, amplifying dependency.

Critically, Gordon-Tapiero argues that emotional dependency and addiction should not be dismissed as side effects but may be viewed as product defects under current consumer law. This could open the door to regulation by litigation—a tool often overlooked. The debate is heightened by leaked internal Microsoft documents revealing plans to 'make people addicted' to its AI assistant. The piece questions how to balance profitability and user wellbeing without hindering innovation, especially as vulnerable users like children are most at risk.