3 ms·
Obviously I used an extreme example, but we shouldn't let machines make decisions that could easily be overruled in an argument with a human arbiter. How we de
by DrSiemer 5y ago
Obviously I used an extreme example, but we shouldn't let machines make decisions that could easily be overruled in an argument with a human arbiter.
How we deal with known offenders is another discussion, but adding these kinds of preemptive limitations to our hardware is unwanted overreach, that sets a bad precedent. Something about a slippery slope.
- elzbardico 5y agoYes, I don't want ML deciding I am not fit to drive tomorrow, maybe I am not fit to have custody of my kids in one week according to the model. But, detecting etanol levels in blood is a fairly well established set of processes that rely on well understood electrochemical reactions. I don't see too much risk of arbitrary unfair decisions from a dumb interlocking device.