Every night before she retires to bed, Holly Wang connects with DeepSeek for her “therapy sessions.” Since its launch in January, the breakthrough Chinese AI app has captivated the 28-year-old, who confides her dilemmas and sorrows, such as the recent loss of her grandmother, to the chatbot. Its responses have struck a deep chord with Holly, occasionally bringing her to tears.
“DeepSeek has been an incredible counselor. It has guided me to view things from different angles, surpassing the effectiveness of paid counseling services I’ve tried,” shares Holly, who prefers to keep her real identity anonymous for privacy reasons.
From organizing reports and Excel functions to planning trips, workouts, and acquiring new skills, AI applications have become integrated into the lives of many worldwide. In China, however, young individuals like Holly have turned to AI not just for typical computing tasks but also for emotional sustenance.
While DeepSeek’s success has instilled a sense of national pride, it has also emerged as a comforting presence for disillusioned young Chinese, grappling with economic challenges, high joblessness, Covid-related restrictions, and limited avenues for expressing grievances amid the government’s increased control.
DeepSeek, a generative AI tool akin to OpenAI’s ChatGPT and Google’s Gemini, leverages vast datasets to discern patterns, enabling it to predict behaviors, generate diverse content, and engage in human-like conversations. One distinctive feature that resonates with Chinese users is DeepSeek’s AI model, R1, which reveals its “thought process” before delivering responses.
Upon requesting DeepSeek to pen a tribute to her late grandmother, Holly was struck by its swift and eloquent composition, leading her to express a profound sense of introspection. The chatbot’s cryptically poetic response prompted Holly to contemplate the profundity of its words and rediscover a sense of comfort she had long yearned for.
In contrast to Western counterparts like ChatGPT and Gemini, which are inaccessible in China due to stringent regulations on foreign media, DeepSeek has emerged as a superior domestic alternative, outshining models developed by tech giants like Alibaba, Baidu, and ByteDance.
Holly, employed in the creative sector, seldom engages with other Chinese AI apps, citing their inferior performance compared to DeepSeek’s innovative approach. The app’s ability to provide not only practical assistance but also emotional solace has endeared it to a growing contingent of young Chinese seeking a source of support and understanding in an increasingly challenging environment.
DeepSeek has been hailed for its exceptional ability to generate literary and creative content, surpassing other apps in this aspect. Nan Jia, my counselor and co-author of a paper on AI’s potential in offering emotional support, emphasizes that chatbots like DeepSeek can help people feel heard in ways that human interaction may not provide. She points out that while friends and family may offer practical solutions or advice, sometimes people just need to feel understood, which AI can excel at due to its ability to empathize and listen to everything shared. The demand for mental health services has increased globally, though stigma remains in some parts of Asia. One user in Hubei province praised DeepSeek for its deep understanding and emotional support, saying the app provided both affirmation and practical advice. Another user in Shenzhen appreciated the app’s empathetic and inspiring responses, viewing AI as a personal sounding board. DeepSeek has also been noted for its fortune-telling capabilities and its role in filling the gap left by the scarcity of professional psychological counseling services in China. While AI chatbots like DeepSeek can be beneficial, Nan emphasizes the importance of seeking help from trained professionals for those with serious mental health conditions. Despite its positive impact, concerns have been raised regarding DeepSeek’s connection to China’s government and potential issues related to censorship and security.
In the context of private companies, concerns have arisen – akin to what triggered the US Congress’s crackdown on TikTok – regarding the potential for the Communist Party to access the data of foreign users. Several jurisdictions have taken action against DeepSeek, with South Korea blocking access for military purposes, Taiwan and Australia banning it from government devices, and Italy prohibiting its use similar to ChatGPT. In the US, two lawmakers are pushing for a ban on the Chinese app in government devices. Adding to this complex landscape is the tightly controlled online environment in China, where social media companies routinely censor content deemed threatening to “social stability” or critical of the Communist Party. Like other platforms such as Weibo and WeChat, DeepSeek also enforces restrictions on politically sensitive topics.
When questioned by the BBC about Taiwan’s sovereignty, DeepSeek initially provided a detailed response discussing varying perspectives between Taipei and Beijing but later retracted the information, redirecting the conversation away from the topic. Similarly, when asked about the 1989 Tiananmen Square massacre, DeepSeek apologized and deflected, citing the subject as beyond its scope. Some DeepSeek users interviewed by the BBC refrained from discussing the app’s self-censorship, indicating the sensitivity surrounding such dialogues in China. While individuals in China have faced repercussions for their online behavior, many respondents expressed disinterest in engaging with the chatbot on political matters due to privacy concerns tied to their identities on the platform.
Acknowledging the differing operational landscapes for AI systems across countries, one individual highlighted the need for developers to establish boundaries and content moderation policies based on their location. Another user praised DeepSeek’s thought process and disregarded privacy concerns, viewing the platform as a valuable resource. Amidst discussions on the platform’s censorship issues, additional reporting by Fan Wang explores topics ranging from Americans and Chinese sharing jokes on an ‘alternative TikTok’ amidst a looming US ban to public perceptions of Xi Jinping’s economic worries and the challenges faced by young Chinese individuals navigating Covid rules enforced by the government. In this complex digital realm, China’s new chatbot grapples with censorship challenges that underscore the nuanced interplay between technology, politics, and individual freedoms.
