Autonomous Vehicles Raising Doubts

Issue 74

Promises of self-driving cars from leading engineers such as Elon Musk may have been misleading as skeptics are now unsure when self-driving cars will reach consumers.  With over $80 billion already invested into the industry, some are concerned that fully functional autonomous cars could take decades to create or may never become a reality.
         
In order for self-driving cars to function properly, scientists believe that the AI controlling the vehicle requires a similar intelligence level to that of humans, however, according to Mary Cummings, a professor of computer science and director of the Humans and Autonomy Lab at Duke University, who has advised the Department of Defense on AI, says that current AI technology has reached only the most primitive forms of intelligence.  Self-driving cars are able to operate given optimum geography and ideal weather conditions, but as soon as extraneous variables are entered into the mix, such as an unmarked construction site, autonomous vehicles quickly fall short.

Alphabet’s subsidiary Waymo has been testing a self-driving taxi in the Phoenix area.  These taxis currently require human supervision from the back seat but still face the same problems when it comes to their AI.  These taxis have a hard time maneuvering around construction zones and have been reported to get rear-ended more often than a human driver due to their confusing, non-human, behavior.

Some argue that altering the infrastructure and adding sensors and transponders can assist the AI and allow for it to not require a high level of intelligence.  Unfortunately, even with an updated infrastructure self-driving car would be limited to specific areas and require certain weather conditions.

While advancements in AI technology continue to bring us closer to the goal of self-driving cars, current technology is far from what is required to bring fully autonomous vehicles to any region of the world.



Source: Wall Street Journal