[THE VIEW]AI Automatic Dispatch Reduces Accidents

27% Fewer Accidents Compared to General Dispatch
Helps Riders Focus on Driving
Initial Inconveniences Must Be Addressed

The introduction of artificial intelligence (AI) has become a hot topic of debate in platform businesses. Especially in transportation and delivery platforms, AI is being actively utilized to efficiently match the rapidly increasing demand and supply, with AI automatic dispatch systems playing a crucial role.

It has already been proven through research that AI automatic dispatch systems can significantly improve platform operation efficiency and rider earnings. In a simulation of AI algorithm-based taxi dispatch in New York City conducted by an MIT research team, the demand that previously required 13,000 taxis was met with only 3,000 vehicles through AI automatic dispatch, satisfying 98% of taxi demand. Furthermore, AI automatic dispatch achieved optimal dispatch without empty vehicles, increasing both revenue and utilization rates. Using AI automatic dispatch also reduces passenger waiting time to about 2.7 minutes, enhancing customer satisfaction. This represents a revolutionary improvement for matching platforms that must meet immediate demand. Based on these advantages, leading overseas platforms are already actively utilizing AI. For example, Uber, a representative taxi application in the United States, operates a system that optimally matches vehicles and passengers through AI. Additionally, the food delivery platform DoorDash uses AI to assign optimal deliveries to couriers.

In Korea, despite this global trend, the adoption of AI automatic dispatch systems is much slower compared to other countries, and debates and resistance surrounding the introduction of the technology continue. In particular, riders are concerned that accepting or rejecting AI-provided dispatches may distract them from driving, increasing the risk of accidents. Consequently, there are worries that AI automatic dispatch systems could negatively affect rider safety and earnings compared to the traditional method where riders select orders themselves. These concerns are understandable.

However, contrary to these concerns, scientific research results suggest that AI automatic dispatch systems can actually help reduce accidents and improve operational efficiency. A joint study by Baedal Minjok and the National University of Singapore research teams found that riders using AI automatic dispatch had a significantly lower probability of accidents compared to those using general dispatch systems. Even after controlling for external factors such as rider gender, age, past delivery behavior, and weather, AI automatic dispatch was shown to reduce accidents by about 27% compared to general dispatch.

AI dispatch systems help riders focus on driving and can effectively reduce cognitive load. The process of accepting or rejecting AI-recommended dispatches offers a much simpler decision-making process than general dispatch, which requires comparing multiple orders by distance, delivery cost, location, and choosing one. Moreover, in general dispatch, riders must compete with others, often needing to continuously check the app and attempt multiple selections. Furthermore, AI systems can quickly perform complex calculations that are difficult for humans, making it more likely to provide better-suited dispatches than those chosen directly by riders. AI has the capability to make optimal decisions by considering past delivery behavior, current location, and even restaurant locations, easily surpassing human limitations in this regard.

AI automatic dispatch systems have the potential to simultaneously improve efficiency and safety, and many global companies are already actively utilizing them with great success. Above all, the spread of AI dispatch systems is essential for safety. Efforts to maximize the benefits of AI must continue from both platforms and workers. Of course, riders may experience discomfort during the initial introduction of AI, and these inconveniences and issues should be gradually improved. However, unconditional opposition and rejection of AI are undesirable. It is important to understand proven facts and make judgments based on scientific evidence rather than relying solely on experience and feelings. Platforms should also actively communicate with riders and provide educational programs to clearly convey the positive impacts of AI automatic dispatch systems on increased earnings and reduced accidents. Through this, riders’ resistance to AI can be lowered, helping them adapt more easily to the system’s introduction.

Kyung Na Kyung, Professor, Department of Computer Science, National University of Singapore