US tech giant OpenAI introduced a ChatGPT-powered tool called “Deep Research” on Monday, just ahead of key meetings in Tokyo, as competition in the AI sector intensifies with China’s DeepSeek chatbot. During a live-streamed announcement, OpenAI researchers showcased how the new agent could analyze web search data to suggest ski equipment for a winter trip to Japan. The company stated that the model could process and synthesize information from hundreds of online sources at an analyst-level capacity.
The demonstration sparked excitement across Silicon Valley. OpenAI asserted that the new model, named DeepSeek-R1, matched the performance of one of its existing models while being significantly more cost-effective. This development reinforces concerns among American AI experts that China’s increasingly powerful AI models are fueling a high-stakes technological race with the United States.
Those worries have sparked a big selloff in the market for semiconductors used to run AI models. In the latest trading session, chipmakers, including Nvidia and Micron, fell sharply. That dragged down the broader stock market, as technology stocks represent a large chunk of the S&P 500.
But not everyone is convinced that the DeepSeek buzz is overblown. Analysts say the tech industry’s reaction seems primarily driven by politics and fears of losing ground to China. Some investors fear that an ultrapowerful Chinese AI model could lead to the overregulation of the sector, which many leaders of US AI companies have warned against.
Another concern is how the Chinese model could be abused, especially by the government. The Chinese company has promised to keep user data private, but it uses a slew of input data to traintostrainThat including the user’s IP address and keystroke patterns. It also stores a database of questions and answers on a secure server.
Some worry that the data collected by the Chinese company could be used to attack the United States and its allies. They point out that the company’s founder is a prominent member of the Chinese Communist Party and that it appears to have close ties to the country’s military establishment.
The attention on DeepSeek is also raising concerns about China’s role in the world. Several American analysts have pointed out that the creators of the AI system are linked to China’s intelligence agency. They have also accused China of using state-owned companies to steal secrets and push back against American tech firms.
DeepSeek is a formidable challenger to the top free chatbot in Apple’s US app store, but it isn’t yet the clear winner despite its rapid rise. The model has yet to face a thorough comparison to rivals like Microsoft’s Skype and Facebook’s WhatsApp. A review of its older model by information reliability analytics firm NewsGuard found that it was only accurate 83% of the time. Moreover, it still has significant shortcomings in understanding and interpreting natural language. The Chinese firm may be relying on outdated methods to build its models, such as reinforcement learning, a notoriously error-prone method for building AI systems.