iEthics: A crush on crashes

Apple wrote a letter to National Highway Traffic Safety Administration regarding its Project Titan which covers its automated vehicles policies. The project is all about automated transportation and smart cities, but maybe not in the way it looks like at first sight. There’s no iCars or iBuses involved, not a single wheel spinning under some fancy aluminium bodied gyro-stabilized urban startup fugitive sport-toy. But there’s some coding and data collecting involved. The company insists on an industry wide cooperation between the main players like Ford and Tesla and Google, on collecting and sharing crash, and near-miss crash traffic data with each other. The letter implies three main areas which could be addressed more effectively, given its suggestions pass through, and these three are:

(1) the implications of algorithmic decisions for the safety, mobility, and legality of automated vehicles and their occupants; (2) the challenge of ensuring privacy and security in the design of automated vehicles; and (3) the impact of automated vehicles onthe public good, including their consequences for employment and public spaces.”

It’s quite obvious, what’s rolling on here: in automated vehicle traffic, there is an enormous quantity of automated decisions made by AVAI (Automated Software Artificial Intelligence) software . Sometimes these are quite simple ones, like stop at a red lamp, or restart an engine. At other times they can be more complex, like should the given car schedule a refill of electricity or gas, and what’s the best route to the nearest refill station. And there are those unpredictable, ugly times, when AVAI will have to decide, whether to hit the little boy suddenly running through the street, or steer the wheel, to the concrete slab 3 foot closer to the bumpers, seriously harming the travelers inside the car.

Although it sounds unsolvable, and our mind furiously trying to work around the question, by solving it with saving everyone, there will be times for sure, when it will be impossible. Just think about the sheer amount of traffic situations happening all at once in every moment of a day in a big city. Well how about a metropolis? And how about a couple of hundreds of them? Its easily the single most complicated issue of the whole automated vehicle industry. And it needs a good and ethical solution, preceded by a dark and difficult debate never seen before, opening discussion topics on the very verge of social Darwinism, aiming to upgrade high speed steel burdened computers with a license to kill. Who, and by which laws will program the iEthics of Judge Idared: yet, noone knows.

At least, the Feds has a Federal Automated Vehicles Policy  which is something to start with. And however thins turn out, chances are, that it will still be safer net of roads, than it was before, when we chauffeured.

Viktor Justin

reflecting on the news at: