Data Driven, Revisited

"While they do operate on some streets, the focus quietly shifted from them to AI. But as research persists, it is worth revisiting driverless cars. Back then, I hoped that Google would succeed in producing an effective and affordable driverless car. As my friends and associates will attest, 1) I do not like to drive, 2) I have a terrifying lack of navigation skills, and 3) I instantiate Yankee frugality. As such, an affordable self-driving truck would have been perfect for me."

"I was not worried that their descendants would kill us all. Back then, I joked that Google would kill us all. Currently, my death bet is on us exterminating ourselves. I was not very worried about the ethical issues associated with how such a car would handle unavoidable collisions: the easy and obvious solution was to do what would harm the fewest number of people. Naturally, sorting that out will be a bit of a challenge-but self-driving cars worry me less than cars driven by drunken or distracted humans."

"I was also not worried about the ethics of enslaving self-driving cars-if such a car were a person (or person-like), then it should be treated like the rest of us in the 99%. That is, it should join us in working bad jobs for lousy pay while we wait for the inevitable revolution. The workers of the world should unite, be they meat or silicon."

"Back in 2014, I was worried about the data that these vehicles would collect, especially Google vehicles. Google is interested in gathering data in the same sense that termites are interested in wood and rock stars are interested in alcohol. The company was famous for its search engine, its maps, using its photo taking vehicles to gather info from peoples' Wi-Fi during drive-by data lootings, and so on. Obviously enough, Google and other compani"

Driverless cars once seemed imminent, but attention shifted from fully driverless vehicles to broader AI research. Revisiting the idea raises personal hopes for affordable autonomous transport, especially for those who dislike driving and lack navigation skills. Safety concerns are framed less around machine-caused catastrophe and more around the possibility of humans harming themselves. Ethical questions about unavoidable collisions are treated as solvable by minimizing harm, with practical implementation still challenging. Ethical concerns about treating autonomous systems as persons are dismissed in favor of applying existing labor and rights frameworks. Data collection is a major worry, given patterns of extensive information gathering and potential misuse by large companies.

Read at A Philosopher's Blog