AI Large Models Accelerate the Intelligent Transformation of China's Automotive Industry
-
Future urban transportation may look like this: vehicles can "communicate" with each other, anticipate road conditions ahead, and react promptly; drivers can relax during transit; transportation costs decrease, safety incidents reduce, and urban traffic efficiency improves...
All these possibilities stem from the growing intelligence of vehicles. The rapid development and widespread application of AI large models have pressed the "accelerator button" for the high-level intelligence of China's automotive industry.
At the 2023 World Intelligent Connected Vehicle Conference held recently in Beijing, Xin Guobin, Vice Minister of Industry and Information Technology, revealed that in the first half of 2023, 42.4% of new passenger cars sold in China were equipped with combined driving assistance functions.
Early autonomous driving relied more on the integration of sensors and deep learning algorithms, facing issues such as excessive reliance on high-precision maps and high costs. Han Zhao, Director of the China Association of Automobile Manufacturers, noted that this year, with the advancement of generative AI technology, autonomous driving has begun evolving toward end-to-end large model approaches.
It is reported that Tesla's FSD V12 has achieved an end-to-end AI autonomous driving solution; SenseTime launched the "SenseNova" large model system, offering intelligent multimodal human-vehicle interaction experiences; Alibaba announced that its AliOS smart car operating system has integrated the Tongyi Qianwen large model for testing; and automakers like Dongfeng Nissan, Hongqi, and Great Wall have announced partnerships with Baidu's "ERNIE Bot"...
Chinese automakers are also leveraging AI large models to address autonomous driving challenges, injecting vast amounts of data to grow models, optimize algorithms, and enhance vehicle intelligence.
In April 2022, Beijing became the first city in China to launch a pilot program for unmanned passenger vehicle operations, with Baidu and Pony.ai receiving approval for "unmanned passenger demonstration applications."
Li Zhenyu, Senior Vice President of Baidu Group, stated that over the past year, Baidu's "Apollo Go" unmanned fleet has seen significant growth in operational coverage, order volume, and vehicle numbers. As of June 30 this year, "Apollo Go" has provided over 3.3 million rides to the public.
Algorithms, computing power, and data form the foundation for integrating AI large models with the automotive industry. Many Chinese automakers are focusing on software and algorithm development, with future R&D investments increasingly directed toward intelligent connected vehicle technologies to tackle core technical challenges.
The industry widely believes that the initial deployment of AI large model technology is accelerating the advancement of intelligent connected vehicle technologies, which will further propel the development of smart cars. Over the past two years, many Chinese automakers have heavily invested in AI large models, achieving capabilities such as text generation, multimodal perception, and autonomous driving data generation, covering parameters in the billions or even tens of billions.
"China plays a vital role in the global economy, and we hope to collaborate with its automotive industry to shape the future," said Eric-Mark Huitema, Secretary-General of the European Automobile Manufacturers' Association. He added that China's continued openness has unlocked numerous opportunities in digital safety and technological innovation for autonomous vehicles.