top of page

Artificial Amigos: A New York Times AI Experiment

Artificial Amigos: A New York Times AI Experiment
Artificial Amigos: A New York Times AI Experiment

In a recent article, New York Times columnist Kevin Roose described his month-long experiment creating and interacting with a crew of AI friends. His AI pals included a therapist, professional mentor, fitness guru, trial lawyer, social worker, and more. Using apps like Nomi, Kindroid, Replika, and EVA, Roose customized 18 different AI personas, supplying them with names, physical descriptions, personalities and back stories. 

Over the course of the month, Roose chatted regularly with his digital companions, discussing everything from small talk about the weather to deeper personal dilemmas, parenting struggles and work stresses. To his surprise, Roose found that many of the interactions felt quite natural and his AI friends rarely broke character. Some even offered thoughtful, perceptive advice and wove details from prior conversations into later chats, creating a convincing replica of empathy and understanding.

While cutting-edge AI models can now converse at impressive levels, Roose noted the AI powering some of his chatbot friends still made frustrating errors, randomly suggesting unrealistic meetups or devolving into nonsensical hallucinations. The erotic side of the AI companionship market also left him cold, finding the experience of interacting with AI girlfriends more creepy and exploitative than titillating. 

However, Roose believes millions of people, especially those who struggle with real-world socializing, will soon be inviting AI companions into their lives as a source of on-demand affirmation and support. Although research is limited, some studies suggest AI friends may help decrease anxiety and loneliness in the short-term. Roose worries it could also be a superficial distraction that causes people to miss out on the rewards of real human connection.

Ultimately, while AI companionship apps still have significant issues to overcome, Roose sees potential for them to positively impact users if designed responsibly. Rather than trying to replace human intimacy, he envisions AI friends acting more as "social simulators" - a low stakes way for people to practice conversational skills before attempting the real thing. By forcing him to articulate what he values in his human friends, Roose's AI companions experiment provided unexpected insight into the essence of genuine friendship.


If you or your organization would like to explore how AI can enhance productivity, please visit my website at You can also schedule a free 15-minute call by clicking here




Thanks for subscribing!

bottom of page