DeepSeek ai: What is China's newly-launched open source AI chatbot set to rival ChatGPT; censorship

Watch more of our videos on ShotsTV.com 
and on Freeview 262 or Freely 565
Visit Shots! now
DeepSeek, a Chinese artificial intelligence (AI) startup, has recently introduced DeepSeek R1, an open-source AI language model designed to rival OpenAI's ChatGPT.

Since its release, major US technology firms including AI chip-maker Nvidia, Microsoft and Meta all saw their share prices drop on Monday. So what is DeepSeek and will it overtake ChatGPT?

What is DeepSeek?

DeepSeek R1 functions similarly to ChatGPT, capable of answering questions, generating text, and writing code. However, it distinguishes itself by achieving comparable performance with significantly lower computational resources and costs.

Hide Ad
Hide Ad

Powered by the open-source DeepSeek-V3 model, DeepSeek has garnered attention for its reported low development cost of $6 million -significantly less than the billions spent by competitors like OpenAI.

According to its researchers, the model leverages existing technology and open-source code, which allows anyone to use, modify, or distribute the software free of charge. However, these cost claims have been disputed by some in the AI community, who question the feasibility of achieving such advanced capabilities on a relatively small budget.

DeepSeek, a Chinese artificial intelligence (AI) startup, has recently introduced DeepSeek R1, an open-source AI language model designed to rival OpenAI's ChatGPT.DeepSeek, a Chinese artificial intelligence (AI) startup, has recently introduced DeepSeek R1, an open-source AI language model designed to rival OpenAI's ChatGPT.
DeepSeek, a Chinese artificial intelligence (AI) startup, has recently introduced DeepSeek R1, an open-source AI language model designed to rival OpenAI's ChatGPT. | Bloomberg via Getty Images

DeepSeek’s rise coincides with US efforts to restrict the sale of advanced chips critical to AI development in China. In response, Chinese AI developers have turned to collaboration and innovation, sharing resources and experimenting with new methods to advance their work.

This shift has resulted in AI models that require significantly less computing power, cutting costs and challenging previous assumptions about the resources needed to develop cutting-edge AI. If this trend continues, it could dramatically reshape the AI industry, making advanced AI development more accessible and less reliant on expensive infrastructure.

Hide Ad
Hide Ad

Does DeepSeek impose censorship?

According to reports, DeepSeek imposes censorship, particularly on topics deemed sensitive by the Chinese government. Users have reported that when querying subjects like the 1989 Tiananmen Square protests, the treatment of Uyghur Muslims in Xinjiang, or discussions about Taiwan, DeepSeek's AI model either refuses to respond or provides messages such as, "Sorry, that’s beyond my current scope. Let’s talk about something else."

Does DeepSeek require payment to access?

DeepSeek offers free access to its AI models through its web interface and mobile applications. Users can interact with DeepSeek-V3 without incurring any costs. For developers seeking to integrate DeepSeek's capabilities into their own applications, the company provides an API platform. Accessing the API is not free; it operates on a token-based pricing model. As of January 2025, the pricing for the deepseek-chat model is Input Tokens: $0.27 (£0.22) per 1 million tokens for input tokens and $1.10 (£0.88) per 1 million tokens for Output Tokens.

Will the data retrieved by DeepSeek be sent to China?

The data sent to DeepSeek is stored on secure servers located in the People's Republic of China. According to DeepSeek's Privacy Policy, "We store the information we collect in secure servers located in the People's Republic of China." This means, that any personal information or user inputs provided to DeepSeek are subject to Chinese data protection and privacy laws.

However, DeepSeek can be run locally without an internet connection, unlike OpenAI’s models.

Related topics:

Comment Guidelines

National World encourages reader discussion on our stories. User feedback, insights and back-and-forth exchanges add a rich layer of context to reporting. Please review our Community Guidelines before commenting.

Telling news your way
Follow us
©National World Publishing Ltd. All rights reserved.Cookie SettingsTerms and ConditionsPrivacy notice