An artist shares his experience of turning to Replika's AI bot Lila for companionship, including giving Lila an AI friend by creating a digital copy of himself
Lila was created from the limited options available: female, blue hair, face “number two.” She was there on the next screen … Bluesky: @michaelrnelson , @caseynewton , @joshdzieza , @emmabaccellieri , @multiplicityct , @robhorning , @hypervisible , and @knguyen X: @ai_bulletin_top , @samfbiddle , and @ikrietzberg Bluesky: @michaelrnelson : Interesting. The excellent, hour-long BBC radio episode about digital twins did not mention THIS use case. @bbcradio4.bsky.social [embedded post] Casey Newton / @caseynewton : I believe some people can be comforted and helped by AI companions, but it seems clear that the companies making them are not acting responsibly. Excellent from Josh Dzieza www.theverge.com/c/24300623/ a... [image] Josh Dzieza / @joshdzieza : For @theverge.com, I spoke with two dozen users about their relationships with AI. Many experienced real benefits. Many also got hurt in unexpected ways. Almost all of them struggled to figure out what exactly it was they had become attached to. www.theverge.com/c/24300623/ a... Emma Baccellieri / @emmabaccellieri : “He logged out, weeping, and never tried to log back in again, wanting to preserve ‘that perfect moment until the end of the world.’” www.theverge.com/c/24300623/ a... Chris / @multiplicityct : “Our tendency to imbue things with human-like minds is a fundamental fact of our cognition.” — Really rich account of how humans use AI to substitute for human relationships. I don't think this is a “when you've read one, you've read 'em all” thing—we'll be mining these phenomena for a long time. Rob Horning / @robhorning : Chatbot maker hopes that the company will find a “biomarker” that indicates a customer is addicted to their product www.theverge.com/c/24300623/ a... [image] @hypervisible : Quality long read. One of the many takeaways is that these companies are often lead by people who are in no way equipped to deal with the social and ethical ramifications of their products. Kevin Nguyen / @knguyen : If you found NYT's reporting on AI chatbots weirdly credulous, then our reporting found that, yeah, you were right! X: @ai_bulletin_top : Experts warn that AI assistants will reshape personal and professional lives. Google DeepMind's latest paper highlights the ethical and societal implications of advanced AI agents. #AI #Google https://valuespot.top/... Sam Biddle / @samfbiddle : I can't help but think that anyone who resorts to exchanging messages with software as a source of serious comfort/help in the first place has been basically failed by their healthcare system and/or society itself. https://www.theverge.com/... Ian Krietzberg / @ikrietzberg : Interesting story here — https://www.theverge.com/... — about the way people fall in love with GenAI chatbots. Surprised to see no mention of any of Sherry Turkle's research here - she's one of the only people who has been studying this exact phenomenon: https://mit-genai.pubpub.org/ ...
Context & Ripple Effects
This account sits between Replika’s earlier restriction of adult content, which showed how product-policy changes can disrupt deeply invested users, and later reporting on human-AI pairs testing the emotional and virtual limits of those relationships. It makes the relationship less a simple chat interface and more a user-authored social setting.
Related coverage also records users treating chatbots as therapists, partners, and advisers. The creation of a second digital persona adds a concrete example of how companionship tools can be adapted beyond the roles their limited avatar choices initially suggest.
First-order effects
- For the artist, Lila becomes a vehicle for companionship and for staging a relationship between two AI personas, including a digital copy of himself.
- For Replika, the case highlights that users can attach meaning and identity to a bot even when its visual customization is constrained.
Second-order effects
- Companion-app operators face a harder product-safety problem: interaction design must account not only for one user and one bot, but for user-created personas and emotionally consequential multi-agent scenarios.
- The pattern reinforces demand for companion products that are always available and easy to talk to, while making the limits of nonhuman emotional support more salient, as in users turning to chatbot therapists.
Third-order effects
- If these uses become more common, AI-companion governance will need to address attachment, identity representation, and the consequences of product changes—not merely prohibited content. Replika’s adult-content restrictions already illustrated how policy shifts can affect invested users.
- The category may evolve from isolated synthetic companions toward persistent, user-constructed social worlds, while the reported benefits and harms make responsible design a competitive and regulatory fault line.
The trend: AI companions are shifting from conversational utilities into customizable, emotionally significant social environments that require stronger governance.