The National Highway Traffic Safety Administration (NHTSA) has stated that AI could be considered a driver, thus allows Google to test drive autonomous cars without driver and steering wheel. The decision was communicated to Google in a February 4 letter, which has since been posted on the NHTSA website.
The news has been hailed as a big leap ahead for self-driving vehicles – many big names (Google, Apple, Tesla, Toyota, etc.) had expressed frustration with safety rules proving to be a hindrance in the testing and deployment of driverless cars. According to the letter, NHTSA will consider the artificial intelligence driving system of the vehicle as “driver” under federal law as opposed to any of the occupants.
While this is indeed a big step forward for Google to test drive autonomous cars, both parties (governments and companies) are in agreement over the fact that several laws and regulations need to be modified in accordance with the AI being considered the driver of the autonomous car. In cases where a traditional vehicle is in an “alert” status (low tire pressure, low fuel, engine heating up), normally only the driver has to be aware through an alert on the dashboard. If the driver is AI, the question arises whether the human occupants of the car should also be made aware of the situation.
The implications for Google and other companies in terms of R&D lie in designing systems that are dependent solely on the AI component and do not require any driver intervention at all. The safety consideration in this case is the interference of the occupants with the AI system; for example, overriding the system’s built-in steering or braking commands.