Roosevelt Institute | Cornell University

The Future of Autonomous Vehicles

By Kathy LinPublished February 24, 2017

null
Dramatic advances in science and technology are leading us into an unknown future far beyond the scope of current regulations in place. The technology behind driverless cars are no longer a future fantasy but rather a present reality. However, despite this readily available new technology, these cars are still rare due to lags in regulation. While regulators have endorsed new rapid-developing technologies, they are now tasked with balancing the interests of companies such as Tesla, Google, and Uber, with the growing public concerns of safety and reliability.
By Kathy Lin, 02/24/17

Dramatic advances in science and technology are leading us into an unknown future far beyond the scope of current regulations in place. The technology behind driverless cars are no longer a future fantasy but rather a present reality. However, despite this readily available new technology, these cars are still rare due to lags in regulation. While regulators have endorsed new rapid-developing technologies, they are now tasked with balancing the interests of companies such as Tesla, Google, and Uber, with the growing public concerns of safety and reliability.
 
There have been many instances of self-driving cars already being tested on the open roads, which may also force the hands of regulators. Tesla already sells semi-autonomous cars with the an "Autopilot" feature,  and Uber has run trials in Pittsburg that allowed app users to order rides from driverless cars. Google has also been testing its autonomous vehicles in California where other tech giants in Silicon Valley (including Apple) have been exploring similar technologies.
 
Within the past year, legislators have made strides in regulating autonomous vehicles. Florida and Michigan have both passed laws that allow the operation of autonomous vehicles without a steering wheel, brake pedal, or even a human driver. In fact, Michigan has passed laws allowing driverless vehicles on state roadways, after previously limiting them to test drives. According to Kirk T. Steudle, director of the Michigan Department of Transportation, the law "allows autonomous vehicles on any road at any time for any reason". This means that as soon as these vehicles go to market, consumers will be able to buy and operate them in Michigan. New laws have also been passed to hold companies legally responsible for their autonomous vehicles' traffic violations, instead of the humans "driving" the vehicle. In contrast, some states only permit driverless cars on public roads in carefully controlled tests with humans in the driver seat. Some states continue to grapple with the definition of "autonomous" and have only authorized research on the potential impact of self-driving vehicles.
 
Regulators want to avoid this patchwork of state laws. The federal government has issued a guidelines for the assessment of passenger safety in driverless vehicles, including sections on crash protection, digital security, and passenger privacy. However, these safety requirements are not nearly as stringent as regulations that are imposed on standard, human-driven automobiles. Much of the vague language used in the legislation corresponds with underdeveloped aspects of drieverless cars — the guidelines were intentionally designed to remain flexible so that it can evolve with new advances.
 
Preliminary data from the National Safety Council estimates that 40,000 individuals died in motor vehicle crashes in 2016, a 6% rise from 2015 and a 14% increase in deaths since 2014. Self-driving vehicles have the potential to improve driver and passenger safety by preventing human error, and can create less congested, less polluted roads. We are currently sitting on potentially life saving technologies which makes the roles of regulators even harder: they must strike a balance between ensuring safety without unnecessarily delaying this lifesaving technology, A federal agency, the Food and Drug Administration, confronts similar issues in the pharmaceutical industry through the drug approval process. Carnegie Mellon University artificial intelligence ethics experts David Danks and Alex John London argue that the regulatory system for autonomous vehicles should be modeled similarly after how the FDA regulates the drug approval process. This model would involve "pre-clinical trials" where vehicles would be tested in simulated environments followed by "in-human" studies into real world environments. This would lead to permit-based testing and, eventually, approval. The model also would involve heavy post-approval monitoring. By formalizing the research and development process of autonomous vehicles, regulators will have more insight (due to greater transparency) and will also be able to assure the public of the safety of these vehicles.
 
These revolutionary technological advances demand innovation in public policy and will require new and different institutional arrangements in our society and government. As autonomous vehicles grow in popularity and become more present in society, their success will be dependent on the public's faith in innovation and the progress made through legislation. Regulators, and we as a society, need to explore new ways of monitoring and shaping the development of autonomous vehicles in order to protect the public from a life with self-driving cars and a life without.