DeepSeek and ChatGPT tackle sensitive questions about China differently

As Chinese AI models navigate political sensitivity, distinct approaches arise

We may earn a commission from links on this page.
Image for article titled DeepSeek and ChatGPT tackle sensitive questions about China differently
Illustration: Justin Sullivan (Getty Images)

This story incorporates reporting from GIZBOT, news24online and The Associated Press on MSN.com.

DeepSeek, a new AI chatbot created in Hangzhou, China, and ChatGPT, developed by OpenAI, are facing scrutiny for their differing approaches to handling politically sensitive topics about China. Released in 2023, DeepSeek has positioned itself as a cost-effective alternative with a focus on technical applications, while ChatGPT continues to advance its reputation for general conversational use. The contrasting capabilities—a result of their underlying architectures—highlight varied responses when addressing questions on sensitive political matters.

DeepSeek’s AI model, dubbed R1, is a behemoth with around 670 billion parameters, making it the largest open-source large language model available. Despite this, DeepSeek follows a broader trend observed in many Chinese AI models, such as Baidu’s Ernie, by avoiding responses to politically sensitive issues. This design choice reflects the regulatory environment within China, where discussion on specific political topics can be restricted. The emphasis on technical problem-solving and logic-based tasks make DeepSeek particularly effective in fields like coding and mathematics.

Advertisement

On the other hand, OpenAI’s ChatGPT employs diverse data sources to offer nuanced and varied responses, a feature that can be applied to questions about sensitive political topics, including those regarding China. This more open-ended design philosophy permits a wider range of responses, informed by a general corpus of knowledge that extends beyond technical applications. While this makes ChatGPT versatile for everyday queries, its approach can sometimes provoke controversy when answers are interpreted as politically charged or sensitive.

Advertisement

The distinction between DeepSeek and ChatGPT underscores the broader competitive dynamics between China and the U.S. in the AI sector. DeepSeek’s emergence, particularly its efficiency in technical applications, has sparked industry concern, as evidenced by significant market reactions like Nvidia’s stock dip. The potential for DeepSeek to disrupt existing models with its cost-effective approach suggests changes in demand for AI development hardware, affecting the landscape considerably.

Advertisement

For users torn between these platforms, the choice largely hinges on individual priorities. Those seeking a tool for technical precision may find DeepSeek more aligned with their needs, given its strengths in areas like coding. Conversely, expecting a well-rounded conversational AI with the flexibility to address a variety of topics would likely favor ChatGPT. As both models continue to evolve, their divergent approaches to sensitive topics will remain a focal point of comparison.

Quartz Intelligence Newsroom uses generative artificial intelligence to report on business trends. This is the first phase of an experimental new version of reporting. While we strive for accuracy and timeliness, due to the experimental nature of this technology we cannot guarantee that we’ll always be successful in that regard. If you see errors in this article, please let us know at qi@qz.com.