3 ms·
The problem with measuring the "harm" of something like AI vs e.g. lead is that with lead there are direct first order chemical harmful effects that are indispu
by ironSkillet 4y ago
The problem with measuring the "harm" of something like AI vs e.g. lead is that with lead there are direct first order chemical harmful effects that are indisputable. With AI, any harm will get expressed via complex social dynamics, which involve so many other variables as well.
- revolvingocelot 4y agoWhich can be written off as laziness or just-world fallaciousness! Even consequences like "first order chemical harmful effects" are mostly ignored by the great-great-grandparent's "economists, lawyers etc"; the way AI will help concentrate money and power in the already overstuffed upper classes will barely be acknowledged.
- rsynnott 4y agoOkay, so for leaded petrol substitute mortgage backed securities. The person who invented those was not going “buahahah, fly my pretties, cause a global financial crisis”; they probably thought they were doing a _good, ultimately of benefit to society thing_. And yet…
- ironSkillet 4y agoAnd therefore we also fail as a society to effectively regulate various complex financial instruments before they actually cause an obvious problem. My point is not that we should or should not regulate AI. It's just that I believe the causal link from certain types of AI usage to concrete harm for society is too complex for enough humans to band together and achieve consensus a priori on regulation.