Multiple youth in China are turning to the AI model R1, launched by DeepSeek in January, to get emotional support. Holly Wang, a Guangzhou resident who works in the creative industry, told BBC that she logs on to DeepSeek every night before going to bed to get “therapy sessions”.
Wang had recently lost her grandmother and had initially turned to the chatbot for help in writing a tribute for her. She later shared a conversation she had with R1 to show how it “outperformed” rival AI bots.
Reading the tribute written by R1 for Wang’s grandmother, she said, “You write so well, it makes me feel lost. I feel I’m in an existential crisis.” Responding to her, DeepSeek’s AI said, “Remember that all these words that make you shiver merely echo those that have long existed in your soul.”
Also read: DeepSeek’s AI models boost Chinese chipmakers’ competitiveness against US rivals
Reflecting on this exchange, Holly told BBC, “I don’t know why I teared up reading this. Perhaps because it’s been a long, long time since I received such comfort in real life.”
Rival AI apps built in the West like ChatGPT and Grok are not available to users in China. They have to pay for a VPN service to access them. However, even as compared to AI models developed by other Chinese companies like Alibaba, Baidu and ByteDance, Wang said, “They are not that great.”
University of South California Professor Nan Jia, who has co-authored a paper on AI’s potential in offering emotional support, said, “AI appears to be better able to empathise than human experts also because they ‘hear’ everything we share, unlike humans to whom we sometimes ask, ‘Are you actually hearing me?’”
Also read: DeepSeek has democratised the AI field. What now?
Another Chinese woman said that other AI apps’ experience has “ended in disappointment” for her but DeepSeek’s R1 AI model has “amazed” her. “When I first read [DeepSeek’s R1’s] thought process, I felt so moved that I cried,” she added.
Sharing further on her experience of using R1, the woman claimed, “DeepSeek has introduced new perspectives that have freed me.” She recalled that she asked the chatbot if oversharing was a “psychological condition” that she should address.
R1’s thought process before answering the woman’s question read, “My response should offer practical advice while being empathetic.” It eventually replied saying the woman’s question was “insightful” and said it wasn’t necessarily a problem. “It may reflect a certain communication style,” the chatbot told the woman on her issues with oversharing.
Leave a Reply