@da1l616
If a self-driving car were to drive of a bridge to avoid crashing into a school bus, most people would propably agree with the manufacturers ethics rules. And also stop buying those cars.
@beckettmazeau2224
Was looking away brushing my teeth when the “Slavery, for example, is a natural part of society” came on and I spit up my toothpaste
@matthewboyd8689
That bike example actually is something that real people do but unconsciously If you are wearing a helmet, people assume that you are safer and therefore give you less room, therefore being more likely to kill you This is why in the Netherlands and a lot of other places they don't have a law forcing you to wear a helmet Although in many of those places as well they have it so that bikes are prioritized and go behind buildings or though parks rather than next to cars
@Livenewme
Guys i solved the "who the ai should save first problem" we develop a system that values people based on the good or bad they've done in their life and credit them accordingly. With this credit system, we can then use this to also see if some people can apply for loans and other things too. 😂
@codykrueger796
This is exaxtly what we discussed in our software engineering ethics class.
@Dexter01992
We all know how's it gonna be: AI will at least at first decide who lives and who dies based on how much money that person has, regardless of how much useful such person is to the actual community. Because that's how its masters want it to be...
@RaphaelRoyoReece
I think you are underselling the adaptability of utilitarianism as the rules of society have their own value - but love the channel and the topics!
@jonathanvilario5402
This is really insightful
@samito8742
tbh in such dangerous situations humans tend to opt out for the simplests solutions due to the fight or flight instincts kicking in (which often times makes room for a lot of errors and mistakes). with that being said maybe AI firefighters shouldn't "decide" on who to save and instead go and save ppl one by one regardless of age, sex, gender and so on.
@m.i.c.h.o
This channel is as if sam o nella and exurb1a had a crossover. Love this stuff
@sino_diogenes
3:00 I think this is an unfair description of utilitarian ethics. It's entirely possible to have a utilitarian calculus where certain lives are weighed more than others. Or you could say that Any human is worth more than any non-human, and both are worth more than any number of "robot lives" (since it's only practical for the robot to self-preserve)
@joostvhts
King is back again! And with the recognisable hat!
@codykrueger796
Humanity First! Yeah!
@kiotw1243
the problem with making a super AI "alligned with humans" is that noone agrees on what's the best for humanity. that's why politics are and always will be a mess.
@minuteowl4421
On the subject of utilitarianism, and the happy people being more valuable than depressed people. I think something that may have been overlooked it projected happiness in the future. Obviously we'd never have exact numbers, but depressed people can turn their lives around, and I think that's something we could roughly account for in terms of AI ethics. I do think utilitarianism (emphasizing people) is the best way to guide our decisions, it's just if we pursue that in policies we have to make doubly sure out bases are covered if we don't want to make serious mistakes. Great video BTW
@Zomdao
I think the entire reason for autonomous robots/vehicles to exist is to avoid such scenarios in which we have to choose. A human with average reaction time may have to decide to hit that school bus or drive off the bridge. But an AI which can drive nearly perfectly will most likely just be able to break in order to avoid hitting that school bus. And having a sufficient amount of firefighting robotd will also ensure no person or hanster will have to be left behind. Though realistically, this is a very optimistic goal, and even if we can reach this, we would still have to start with autonomous machines which will have to make ethical decisions. An automomous car would most likely only be able to effectively factor in every possibility of failure if every car surrounding it is entirely predictable, which they won't be as long as humans are driving them. The same way many firefighting robots will perhaps suffice to save everyone in a building, there will be a time where we have firefighting robots but can't supply enough to ensure perfect safety for each building.
@TobyTopF
Least rabid anti-utilitarianism... While an interesting video and enjoyable commentary, you make a lot of strawman arguments that heavily undermine your point. Your other videos feel a lot more carefully considered than this.
@Avg_Janitor
Good stuff
@ab562
This is the best channel on Youtube. (Obviously I am a human)
@MisterZimbabwe