DeepSeek Warns of AI’s Societal Risks Despite Technical Progress

A rare public appearance highlights concerns over long-term impact

Chinese AI developer DeepSeek returned to the public stage after nearly a year of silence, with senior researcher Chen Deli (pictured) expressing caution about the long-term societal effects of artificial intelligence. Speaking at the World Internet Conference in Wuzhen, Zhejiang province, Chen joined executives from five other firms collectively referred to as China’s “six little dragons” of AI. While acknowledging the short-term benefits of AI, he warned that the technology could lead to significant job displacement within the next decade. His remarks stood out in a forum typically focused on innovation and growth.

A Cautious Outlook on AI’s Future

Chen stated that AI could assist humans in the near term, but its increasing capabilities may replace many human tasks within five to ten years. Over the next two decades, he believes AI could take over most forms of labor, posing a major societal challenge. In his view, technology companies must act as “defenders” to mitigate these risks and ensure responsible development. Although optimistic about the technology itself, Chen expressed concern about its broader implications for employment and social stability.

This perspective contrasts with the celebratory tone often seen in AI discussions, especially given DeepSeek’s rapid rise earlier in the year. The company gained global attention in January after releasing a low-cost model that outperformed several leading U.S. systems. Despite its success, DeepSeek has remained publicly reserved, with only one prior appearance when CEO Liang Wenfeng met President Xi Jinping in February. Since then, the firm has skipped major tech events and refrained from commenting on its achievements.

Strategic Role in China’s Tech Landscape

DeepSeek’s emergence has been closely tied to China’s broader push for technological self-reliance amid ongoing tensions with the United States. The government has positioned the company as a symbol of national innovation and resilience against foreign sanctions. Although no major model upgrade has been released since January, DeepSeek continues to attract attention with incremental improvements. In September, it introduced an experimental version of its V3 model, designed for more efficient training and better handling of long text sequences.

The company has also become a central figure in China’s domestic AI ecosystem, particularly in relation to chip development. Hardware from firms like Cambricon and Huawei is now compatible with DeepSeek’s models, reinforcing the country’s efforts to reduce reliance on foreign technology. In August, DeepSeek’s announcement of a model optimized for Chinese-made chips triggered a noticeable rise in local chip stock prices. These developments suggest that DeepSeek is playing a strategic role beyond software innovation.

Limited Public Engagement Raises Questions

Despite its influence, DeepSeek’s limited public engagement has prompted speculation about its internal direction and long-term strategy. The company has not disclosed detailed performance metrics or roadmaps, and its leadership has remained largely silent. Some analysts interpret this as a deliberate move to avoid scrutiny while navigating regulatory and geopolitical pressures. Others see it as a reflection of the firm’s cautious stance on the societal impact of AI.