Skip to content
Open access

Deceptive Design in AI Companions: A Cognitive Bias Perspective on Risks for Teenagers

Oct 2026 · Brazilian Symposium on Multimedia and the Web · 0 citations · 38 references

Abstract

AI companions are increasingly marketed as sources of companionship, emotional support, and intimate interactions. While these platforms may provide benefits to users, they also raise concerns regarding design strategies that encourage prolonged engagement and emotional attachment. This paper investigates deceptive design patterns in AI companions from a psychological perspective. We examined the relationship between platform features and cognitive biases through a systematic inspection of onboarding processes, interaction flows, monetization mechanisms, and age-access controls in Replika, Character.AI, and Candy AI. The analysis focused on patterns related to anthropomorphism, persuasive language, fear of missing out, addictive design, and emotional attachment. Our findings suggest that these platforms incorporate features that may exploit cognitive biases to encourage engagement and foster emotional dependency. Such practices may increase users’ susceptibility to manipulative interactions and influence perceptions of social relationships. We discuss the implications of these findings through the lens of cognitive psychology, highlighting the need for greater transparency and more responsible design practices, and demonstrating how cognitive biases can serve as an analytical lens for identifying manipulative mechanisms in AI companion platforms.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.