The future of robotaxi

Tesla’s unveiling of Robotaxi has been postponed by two months from August 8. The Tesla stock plunged as the second quarter earnings were not satisfactory. Robotaxis are self-driving cars with no steering wheel nor pedal. One of the key purposes...

LEE SU-HWAThe author is a research professor at the Center for AI Convergence Research at Hallym University

Tesla’s unveiling of Robotaxi has been postponed by two months from August 8. The Tesla stock plunged as the second quarter earnings were not satisfactory. Robotaxis are self-driving cars with no steering wheel nor pedal. One of the key purposes of Testa’s FSD, or Full Self Driving, is the self-driving taxi service. Hopefully, the car will be making money while its owner does not use it.

Competition in the robotaxi market is fierce. Alphabet’s Waymo is operating a Robotaxi in Phoenix, San Francisco, L.A. and Austin, Texas. Waymo plans to invest an additional 5 billion dollars. Cruise under GM is seeking to commercialize it in San Francisco. China’s Baidu is operating a similar service in Beijing, Shanghai and Guangzhou.

The road traffic situation in large cities, where robotaxis have already appeared, is complicated. Human drivers also have to make new decisions for newly emerging situations. One of the many challenges that require solutions is to make a decision when an unlearned, unrepresented situation occurs.

Intelligence Augmentation (IA) has emerged to help solve the challenge of the decision making process that robotaxis face. IA refers to strengthening human intelligence with artificial intelligence (AI). It analyzes tasks performed by humans in advance to divide roles between AI and humans. For example, if an AI tool, such as the Acceleration Control for Pedal Error (ACPE), is applied to prevent senior drivers from making mistakes on brakes, accidents could be reduced significantly. The Autonomous Emergency Braking System (AEBS) steps in and stops a vehicle if there is a risk of collision as a driver dozes off or drives recklessly. The ACPE and AEBS are far simpler than other obstacles that a robotaxi needs to get through. The AEBS is already required in trucks.

In order to implement IA, experts must collect, organize, classify and analyze human experience and behavior data in various situations, and then conduct machine learning with the data. This series of processes is also called “user experience design (UXD).” If a smartphone’s user experience (UX) provides ease of use, the automobile’s UX determines life and death.

IA is also very important in school education. It becomes easier to develop students’ intelligence by providing AI assistance. AI digital textbooks perform an important task of enhancing students’ intelligence by helping teachers. The first step is collecting behavior data of students. At first, it will be full of difficulties, but when challenges are overcome, the IA by AI for students will be possible.