New Generation AI Empowering the Era of Smart Mobility – Insights from Didi’s Chief Scientist Tang Jian

On April 18, the 2019 International Intelligent Technology Summit, organized by the New Generation Artificial Intelligence Industry Technology Innovation Strategic Alliance, was held in Qingdao. Tang Jian, Chief Scientist of Intelligent Control at Didi Chuxing, delivered a keynote speech titled “New Generation AI Empowering the New Era of Mobility.”

The article is based on an interview with Tang Jian conducted by Yiou.com, with permission from the original author.

According to Tang, the next generation of AI technologies—big‑data intelligence, cross‑media intelligence, hybrid‑augmented intelligence, autonomous intelligence, and collective intelligence—find their optimal application in intelligent transportation. Massive travel data generated by Didi provides the foundation for these technologies.

Key AI‑driven improvements highlighted include:

Big‑data intelligence for extracting useful information from the huge volume of travel data.

Hybrid‑augmented intelligence (human‑machine collaboration) to coordinate drivers with autonomous vehicles.

Autonomous intelligence as the core technology for future driverless cars.

To enhance user experience, Didi has introduced features such as “recommended pick‑up points” and an upgraded intelligent dispatch system that leverages reinforcement learning to achieve optimal driver‑passenger matching while considering future supply‑demand dynamics.

Safety is addressed through a comprehensive AI‑powered workflow: pre‑trip identity verification via facial recognition, real‑time trip sharing, safety monitoring, one‑click emergency alerts, voice recording protection, and post‑trip driver training. These capabilities rely on natural language processing, speech recognition, facial recognition, and cloud computing.

In terms of core AI research, Didi is advancing five foundational technologies—speech recognition, computer vision, natural language processing, knowledge graphs, and reinforcement learning—to continuously improve the travel experience and platform optimization.

Beyond service upgrades, Didi is expanding into intelligent traffic infrastructure, smart vehicles, and shared mobility. The company has optimized traffic signal parameters at over 1,500 intersections in more than 20 Chinese cities, reducing congestion by 10‑20% through real‑time data analysis.

On the vehicle side, Didi’s in‑car solutions employ speech recognition and computer vision for voice commands, driver behavior analysis, and safety assistance. Didi also operates a 100‑person autonomous‑driving team with over 40 test vehicles, holding road‑testing permits in California and Beijing.

Didi processes over 106 TB of new trajectory data daily and handles more than 4,875 TB of total data, providing a massive data advantage for AI development.

Internationally, Didi collaborates with partners in Japan, Australia, Brazil, Mexico, and other regions, positioning itself as a global player in smart mobility.

Overall, the article illustrates how AI, big data, and cloud technologies are reshaping transportation, highlighting both achievements and ongoing challenges such as model size, computational resources, and safety assurance.