At a rare public appearance, Chinese artificial intelligence (AI) startup DeepSeek issued a stark warning: the long-term advance of AI could upend jobs and disrupt social structures, and companies must act by openly disclosing which professions are likely to disappear.
According to reports from the South China Morning Post and STAR Market Daily, leading Chinese tech companies — including DeepSeek, Unitree Robotics, BrainCo, Game Science, ManyCore Tech, and DEEP Robotics — gathered in Wuzhen, Zhejiang, for the World Internet Conference to discuss trends spanning brain-computer interfaces, robotics, spatial intelligence, gaming, and open-source AI.
The most closely watched speaker was Chen Deli, a senior researcher at DeepSeek, whose comments stood out for their philosophical and social undertones. Other participants included Xingxing Wang, founder of Unitree Robotics; Bicheng Han, CEO of BrainCo; Xiaohuang Huang, chairman of ManyCore Tech; Qiuguo Zhu, CEO of DEEP Robotics; and Ji Feng, producer at Game Science.
The three phases of AI and humanity
Chen outlined a three-stage framework describing how technology's relationship with humanity may evolve, and the roles companies must play in each phase.
In the short term (three to five years), he said, AI and humans are entering a "honeymoon period," in which collaboration yields immense benefits. During this stage, companies should act as "preachers", ensuring technology remains open and accessible, a principle behind DeepSeek's commitment to open-source AI.
In the medium term (five to ten years), as AI begins to replace certain professions, Chen argued that companies should become "forewarners", alerting workers to the need for reskilling and transition.
In the long term (ten to twenty years), he warned, AI may supplant most forms of human labor, potentially triggering social instability. At that stage, technology companies must evolve into "protectors", protecting social order and human welfare.
Despite the risks, DeepSeek remains committed to developing artificial general intelligence (AGI). Chen was candid: slowing AI progress is unlikely, he said, because "everyone is trying to profit from it." The ultimate measure of the AI revolution's success, he added, may be "whether it can replace most of the work done by humans."
Echoes of a global concern
Chen's comments mirror a growing global unease about AI's trajectory. Last month, more than a hundred AI experts — including Zhipu AI CEO Peng Zhang, Tsinghua University professor Andrew Chi-Chih Yao, and AI pioneers Geoffrey Hinton and Yoshua Bengio — signed an open letter calling for a pause in the development of superintelligent systems until society reaches a shared understanding of the risks.
Article edited by Jack Wu
