Is this a real issue? — Uneducated Pleb

Of course it is.

A lot of what people are calling "artificial intelligence" is just better programs operating with a lot more data. For instance, speech recognition is not the product of artificial intelligence--which is not to under-rate it. I think Google's speech-to-text service is excellent at times and is improving. But when you reach a computer on the phone, and when it asks you a question, and it says "Say 'yes' or 'no'" and you say "no" and it says, "I'm sorry, I didn't understand that; let's try again" and 10 tries later it still can't understand "yes" or "no", it's clear that one is dealing with a very stupid and very poorly engineered system.

I've done a fair amount of clerical work, and as far as I am concerned, this is stuff that computers should be doing. There are a lot of jobs that people do that will be eliminated in the next 20 years. Attaching RFID (radio Frequency ID) devices to objects sold in supermarkets eliminates the need for checkout. Readers can survey the chips in your bags, and charge you automatically. There's no artificial intelligence in that.

The autonomous automobile or truck either needs specially adapted superstructure (sensors, signals, data streams, etc.) to aid an on-board computer as it travels down the road, or it needs something a lot closer to (or actual) machine intelligence: A machine intelligence that can make sense of chaotic traffic and respond in real time, and make ethical decisions (according to prescribed rules), or both. My guess (my fear) is that the on-board systems, IF they are even better than Google's speech to text systems, will result in a lot of injured or dead people.

Is this a real issue? Or is it a manufactured outcome from a manufactured problem that will never occur in reality and being bent to political purposes? Somewhere in between? Thoughts? — Uneducated Pleb

It's difficult to say. I veer back and forth on this topic, sometimes I read things that make the AI/robot revolution seem imminent, in which case yes, I think it's a huge potential danger, economically and in many other possible ways, and we really should be putting the brakes on and thinking things through. We ought not to just blithely do tech things simply because we can. "Because it's there" is a fine enough excuse for climbing mountains, but I think we need to be a bit more responsible with technology and have a bigger discussion. In that sense, even if AI isn't imminent, we should be having the conversation, and the smart bods should be thinking through all the possibilities they can in relation to ethical AI, economic ramifications, etc.

We need to do our best to survey the possibility space before we careen headlong into it.

On the other hand, sometimes I read things that make me think the claims for AI are still vastly over-inflated, just as they've always been, and that we're actually nowhere near what the claims would make us think is just round the corner.

Sometimes I think a mix of both - maybe it's coming, but maybe opposition to it is as short-sighted as Luddism was in its day; on the other hand, maybe this is the one time where Luddism is actually the appropriate reaction.

At the moment, I'm in a place where I think that machine learning isn't actually intelligence, so we're still in the area of more and more sophisticated expert systems that are competent but not comprehending, in which case all the danger is still with the programmers/users, not with the AIs themselves or what they're intrinsically capable of.

I think intelligence has much more to do with sociality than people think who think of it as something that can be implemented in a single box, so to speak. That said, maybe it's possible that a community of robots, or a community of programs in a single box, might produce intelligent robots or intelligent programs. But I think it does have to be the result of a community of interacting peers in some sense, and it's not something a single thing can come up with or develop by itself. For one thing, I think the possibility of lying is the kernel of what develops into the capacity for reflection, and for that to evolve, there has to be someone to lie to and something at stake (something to gain from lying, something to lose from being found out).

good point about community. One robot would probably not wend its way through complex traffic; if all the robots were linked in a community, however, so that the robots knew what the other near-by robots were planning on doing next, traffic itself would become "intelligent".

You make a number of good points.

The biggest problem of AIs is that they will not be able to deliver on the promise. The second biggest problem (or the biggest, depending) is that AIs will eliminate jobs. Automation, efficiencies, robots, etc. have already eliminated a lot of jobs, and not just in manufacturing. Computers make it possible for fewer people to manage operations.

One solution to the unemployment problem is the Guaranteed Minimum Wage (GMW) which would supply an income to people displaced by automation, smart robots, AIs and so forth. Depending on their age and abilities, they could either retire early (in reduced circumstances) or they could retrain to do something else.

GMW isn't really all that radical an idea.

“That one innovation,” he continued, “will be enough to create riots in the street. And we’re about to do the same thing to retail workers, call center workers, fast-food workers, insurance companies, accounting firms. — Uneducated Pleb

Let me guess what that 'innovation' is ----- lay-offs (not the driver-less trucks).

One robot would probably not wend its way through complex traffic; if all the robots were linked in a community, however, so that the robots knew what the other near-by robots were planning on doing next, traffic itself would become "intelligent". — Bitter Crank

I think that would still be competence, or machine learning, it would still be something the AI was programmed to do. But to me intelligence - of the kind that we might worry about the thing having ideas of its own - has to do with reflection, which means that an intelligent being is situated in a nexus of other beings like it, and it has a sense of how it looks to them, and it has a sense of it mattering how it looks to them.

I base this speculation on the fact that all the most intelligent creatures we know up to and including ourselves are social creatures (with the exception of the octopus, but even the octopus seems to have brain-like structures for each of its tentacles) - apes, wolves, corvids, parrots, etc.