货拉拉公开安全防控及判责算法，AI应用降低风险单量30%

Core Viewpoint - The company has enhanced its AI safety monitoring and liability assessment algorithms to improve safety for freight drivers and streamline the liability experience, following the cancellation of the liability principle in September [1][4]. Group 1: AI Safety Monitoring - The AI safety monitoring system, developed over four years, can detect hazardous situations in real-time, such as dangerous goods transport, illegal passenger transport, fatigue driving, and overloading [1][3]. - After the full application of AI, the daily risk incidents related to hazardous goods transport and illegal passenger transport have decreased by 30%, with over a thousand orders being forcibly canceled daily due to violations [3][4]. Group 2: Liability Assessment - The AI liability assessment algorithm automatically collects various data post-order cancellation, analyzing it against platform rules to preliminarily determine liability, with a focus on ensuring drivers are not held responsible for canceled orders due to violations [1][4]. - The time taken for human-machine combined liability assessment has been reduced from 72 hours to 48 hours, and the appeal rate from drivers has decreased by nearly 30% since the implementation of AI [4][6]. Group 3: Driver Safety and Education - The platform has pushed out 43 safety courses over the past year, with over 17 million completions, and has invested 1.75 billion yuan in providing insurance for drivers, covering 890 million orders [3][4]. - The AI system sends fatigue driving alerts to drivers after four hours of continuous driving, promoting safer driving practices [3][6]. Group 4: Future Developments - The company plans to continue optimizing its algorithms and listening to driver feedback to create a safer and fairer working environment for freight workers [6].