China Using DeepSeek's AI Model To Develop Advanced Warplanes: Report

China is using its homegrown artificial intelligence platform, DeepSeek, to design new warplanes, according to a report in the South China Morning Post. Wang Yongqing, the lead designer at the Shenyang Aircraft Design Institute, a subsidiary of state-run Aviation Industry Corporation of China, confirmed that his team had started using AI tech to develop the country's most advanced warplanes.

"The technology has already shown promising potential for application, providing new ideas and approaches for future aerospace research and development," said Mr Wang, adding that his team was conducting in-depth research into large language models to resolve "complicated problems based on practical needs".

“This is an important step and points to the future direction of aerospace research."

Mr Wang has been working at the institute for four decades which is responsible for designing multiple fighter jets for military, including advanced multirole warplanes such as the navy's J-15 Flying Shark and the J-35 stealth fighter.

As per Mr Wang, AI was also being used to free up the researchers from performing tedious review tasks so that they could focus on more critical work. 

Hangzhou-based DeepSeek has rapidly emerged as a notable player in the global AI landscape, releasing a series of models that compete with Western counterparts while offering lower operational costs.

Also Read | Fake Mount Fuji In China With White-Paint Hilltop Scams Tourists

DeepSeek R2 model

The development comes in the backdrop of DeepSeek preparing to launch DeepSeek-R2, the successor to DeepSeek-R1. It is expected to be cheaper and better, giving tough competition to the likes of ChatGPT, Gemini and Claude among others.

The new advanced model, R2 is said to have been developed with a so-called hybrid mixture-of-experts (MoE) architecture, making it 97.3 per cent cheaper than OpenAI's GPT-4o model. MoE is a machine-learning approach that divides an AI model into separate sub-networks to jointly perform a task.

In March, the company released a major upgrade to its V3 large language model, intensifying competition with US tech leaders like OpenAI and Anthropic.