New AI digital replicas could fully duplicate famous people, raising ethical alarms 

Martin Seligman, an eminent American psychologist, has an AI chatbot designated after him, which is part of a larger wave of AI chatbots modeled after real humans, sparking debate about the potential perils of AI.

 Yukun Zhao, an old graduate student, persuaded the psychologist that he could develop a “virtual Seligman.”

Over the course of two months, Zhao and his team built an eerily accurate version of Seligman himself, a talking chatbot whose answers drew deeply from Seligman’s ideas, whose prose sounded like a folksier version of Seligman’s own speech, and whose wisdom anyone could access.

According to Politico, the bot, dubbed “Ask Martin,” was created by academics in Beijing and Wuhan without Seligman’s approval or even awareness.

The Chinese-built virtual Seligman is part of a larger wave of AI chatbots modeled on real people, which use powerful new algorithms called as big language models to imitate their online personalities. Meta is experimenting with licensed AI celebrity avatars, and there are now numerous online chatbots educated on publicly available information about deceased historical figures.

In a similar vein, in Southern California, software entrepreneur Alex Furmansky scraped podcasts from the internet to develop a chatbot replica of Belgian celebrity psychologist Esther Perel. He utilized the bot to help him cope with a recent heartbreak, recording his experience in a blog post that someone else eventually emailed to Perel herself.

At the 2023 SXSW conference, Perel addressed the existence of AI Perel. She, like Seligman, was astounded rather than upset by the replication of her personality. Politico said that she referred to it as “artificial intimacy.”

Both Seligman and Perel eventually opted to accept the bots rather than fight them. But it’s unclear whether they would have been able to shut off their digital replicas if they had chosen to. 

It is not illegal to train AI on copyrighted works, which is why it’s unclear what the real Martin could have done to prevent access to the false one, a clone trained on his own thoughts, using his same words, to produce all-new responses.