[KaiserJeep]

In the Coming Automated Economy, People Will Work for AI
A new role for humans: prepping data so AI can learn to do our jobs
By Eliza Strickland, 30 Nov 2018 | 15:47 GMT


In Texas, a company called Alegion is helping disabled veterans take part in the new digital economy. The vets’ job: preparing data so that an artificial intelligence (AI) system can learn from it.

“There’s a whole new industry sprouting on the shoulders of AI,” says Alegion CEO Nathaniel Gates in an interview with IEEE Spectrum. “We are employing people.”

When people talk about AI, they’re often referring to software that gets very good at a particular task via a technique called deep learning. With this method, AI systems are given vast amounts of labeled data, and as they run through it, they learn to draw conclusions. For example, an AI tasked with classifying photos of animals would look at millions of images labeled cat, dog, hedgehog, and so forth, and would learn on its own which features define each animal.

The rest of it: https://spectrum.ieee.org/tech-talk/robotics/artificial-intelligence/in-the-coming-automated-economy-people-will-work-for-ai

...and before anybody points this out, YES this is a blog post, not a typical IEEE news article. One woman's opinion, but I thought you would want to read it.

[KaiserJeep]

I see you did your usual trick of not reading the source article completely. Open mouth, insert foot. The AI applications already exist, you interact with many of them online and on the phone every day, using keypads and voice recognition and mouse/keyboard. This article is about temp jobs for people displaced by automation, or taking job training, who need interim income.

As for whether or not this will happen - well, time to wake up, it happened in the past, these are the final details to tweek the utility of the AI.

I found myself choosing to interact with an AI when I deal with my Medicare Supplement Insurance provider. That as it turns out already works better than interacting with live people in Detroit who have heavy Black dialects and do not really care about what they are doing. My insurance provider gets some Kudoes just for having real people who don't live in India or Pakistan. But if I ever find another insurance company that has real Americans who care about doing a good job, I'll switch in a heartbeat.

Meanwhile the AI is good and slowly improving. You see, some people never were worth a damn, and were obsoleted by technology before they even matured and entered the job market. Well, guess what: no matter what, for most people there is a computer going after your job, unless you steadily get better at what you are doing. Many companies routinely monitor everything employees do as part of their jobs, with the purpose of automating that job at some future time.

One way of looking at what is happening NOW is that these humans who are tagging images and correcting voice recording transcripts are Quislings betraying our species to the race that will replace us.

Well, I actually agree with you: the computers will never completely replace humans. But what they have already done is automate over half the jobs that existed in your lifetime, and that percentage is growing every day.

I'm looking at replacing the two buck-toothed morons who keep grazing the flowerbeds at the Nantucket property with their stand-on professional mowers, for example - which I can plainly see on the HD IP cameras of my security system. I can easily buy a robot mower for what I am paying them in a season, to mow and trim. I will have to bury a boundary wire to keep the mower within bounds and out of the flowerbeds, plus install some more mowing strips, which is the only reason those two 20-something surfer bums are still getting any of my money.

[KaiserJeep]

Doom, doom, doom.

[KaiserJeep]

Yeah, well the difference is that at age 67 and unlikely to live beyond 87 or so, I simply don't care a lot any more, because even the "Seneca Cliff" is more than 2 decades long. So I'm thinking less of prepping and more about conserving assets for the kids and grandkids.

And WTH, if I find myself a lonely old man, I can blow it on Booze and Broads.

[evilgenius]

This isn't much different that when I once worked as an office temp, and watched scores of people in that office become forced to train their replacements. The company was moving its regional headquarters from one state to another. They held a one month for every year worked severance package over the heads of the people who were training their replacements in order to get them to do it. If somebody had worked there a long time they would get a lot more than unemployment would give them. Those who hadn't, the company didn't want them training the new people anyway. If you took the deal you wouldn't get unemployment. It was the people's decision whether a little extra money was price enough to suffer the humiliation. Nobody turned it down. They all grumbled.

You know, the problem with something like this isn't just getting people to prep the data. It's also whether the data is being prepped into the right sort of context. That can be as seemingly innocuous as calling a semi truck a semi truck or a big rig. And that's not to start considering all of the different sorts of large transport trucks. Some are smaller and meant for inter-city hauling. Others are meant to do over the road. So, there is a database aspect issue, a developmental one, which has to be overcome not only to begin with, but, most likely, iteratively.

There is also the problem of what companies consider their purpose to be. The won't admit it, but many companies think that their purpose is to make money, or to spin the money that is made off to management. As such it is very difficult to streamline their systems. When it doesn't matter how slow and inefficient the process runs, only that the big boss's targets are being met so that that he can earn his bonus from the corporation, there will be too little input from the working side of things for an analyst, especially an AI one which needs that sort of honesty, to function. Tracking stuff costs money. Spending on that would add to a cost center that, if it came in as too expensive, could have an impact upon a manager's bonus. If upper management isn't committed to AI, then the working structure within the various divisions of the company will see it stifled. And they will have paid those veterans to look at things and classify them the wrong way, with no one to recognize that it was being done wrongly all along because to do so would cost those people who could something.

[KaiserJeep]

The difference is that database tools - meaning the relational databases that underly every search engine and every AI application, are cumulative. So are human minds, of course, but humans always start from zero and learn to maturity, and then get "zeroed out" by death. Then when human knowledge is used to tweek the databases, those databases become more than the sum of one lifetime, they are hundreds and thousands of lifetimes, soon to be millions.

Today, these databases are incredibly detailed but the knowledge they contain must be interpreted for content by humans. But that is what these tagging operations are doing - establishing relational links between otherwise unrelated datums.

I personally, after 35 years designing computer hardware, do not believe in the "singularity". But the knowledge is becoming more accessible to human minds as time passes. I don't know about the rest of you, but I can enjoy searching and stuffing my head with new thoughts for hours at a time. Everybody on Earth with Internet access effectively lives in the world's biggest library.

Even the Archives of the PeakOil Forums are an effort to synthesize new thoughts from the collective minds of the members. Original synthesis abounds here, which is why I visit so frequently.

[mmasters]

A bitch to a machine, a new low lol. Really though data is knowledge is power. This is nothing new. The AI will give some new insight though.

[evilgenius]

I should have said intra-city.

Anyway, I was thinking about the protests in France in relation to this. Those are leaderless. All they need to remain cohesive is a nudge here and there, keeping those who have a common purpose together when life generally separates them. The French protestors come from every walk of life. You know, people can be brought together. It's much harder to keep them there. Capitalism uses marketing. Marketing runs its course. We call that an ad campaign. A really well built social media AI could identify things about people that they may not even be aware of themselves, and nudge them into groups and activities. It could even use the social equivalent of money, something like recognition, to cause them to act in certain ways toward each other. By doing so over time, it could keep them there. People like to be preoccupied. They like it in a sort of way where it doesn't feel like work, and they can at least pretend they are in control. In the coming world where without the old sort of work people may have too much spare time, an AI like that could do that for them. It would be good for people who know ahead of time what they like. It could foster hobbies and book clubs.

Imagine the implications if one country wrought this sort of tech upon another, and the inflicted upon country failed to recognize what was going on. It would seem like legitimate protest. It could remove governments. At the very least it could create a favorable situation for things like trade with the country clandestinely inflicting the action upon the suffering country. It's not far from protesting high fuel prices to accepting dependence upon cheap Russian natural gas, for instance. If you can get them that far it may also be possible to scare them off of the alternatives. I'm not saying that's what's happening. I'm just pointing out it could happen, and it wouldn't look much different than what we are seeing now.

[Cog]

The French people elected exactly the type of government they wanted. But now that government wants them to pay for the socialism that they so desperately desired. Here is your carbon tax you wanted. Reap it.

The yellow jackets get zero sympathy from me.

[evilgenius]

When I was growing up Martin Luther King Jr. was one of my heroes. Nonviolence was the primary reason. It gets to the truth in a much more lasting way. If a person just gets according to their demands without even the capacity for self examination, then the likelihood of that lasting is less. Even when you are a victim, timing is important. It matters to you, as to whether you can integrate what you want into your life along with what you already value, and it matters to those outside of you, as they will only give you charity if you are asking out of turn. Charity only lasts until the next cycle round. In either case the last state of that person is more likely to be worse than the first, if they have not gotten with acknowledgment to the time.

Who is it that does everything right, but winds up at a loss except the one for whom chance was not there? Don't people reap what they sow? But there are those who repeat behaviors out of circumstances within which they are caught. Some people suffer under crooked justice systems. Others will never get that job which could show them the way out. They can suffer in the same way as a city can have a dangerous intersection where too many people die, and no one does anything about it. In that case it is a person's choice whether they will satisfy themselves at their loss with some kind of ad hoc roadside memorial to their fallen loved one, or take the action to the specific causes why nothing is being done. Especially then, for why should your noise amount to more than the eyesore of that memorial unless you know how to direct it, there is a requirement to engage with proper argument and match skill with skill, attending to the time.

[KaiserJeep]

For a current and acurate update on the topic of AI, I reccomend the Amazom Prime documentary Do You Trust this Computer?

https://www.imdb.com/title/tt6152554/

Not to nitpick, but what we have been discussing and what is nibbling away at the human job force is more properly labelled "intelligent systems" than "artificial intelligence".

But then, neither term actually means what the words comprising it seem to mean, either. Machines have something humans don't have that is not intelligence.

Personally, I don't think we have to worry. The sum of the two species of digital intelligent systems and real human intelligence is something greater than either, no matter which is larger.

[careinke]

For a current and acurate update on the topic of AI, I reccomend the Amazom Prime documentary Do You Trust this Computer?

https://www.imdb.com/title/tt6152554/

Not to nitpick, but what we have been discussing and what is nibbling away at the human job force is more properly labelled "intelligent systems" than "artificial intelligence".

But then, neither term actually means what the words comprising it seem to mean, either. Machines have something humans don't have that is not intelligence.

Personally, I don't think we have to worry. The sum of the two species of digital intelligent systems and real human intelligence is something greater than either, no matter which is larger.

I also recently watched it, and recommend it.

I both look forward to the upcoming AI revolution, and at the same time fear it. I don't fear the AI, more the people who will control it.

[evilgenius]

I think people will always be the scariest thing. In a similar way to man's dominance over the animal kingdom, where man is always to be most feared, man will dominate AI. My concern is that dominance could come about at the behest of natural groupings of people. As such, those groups could sometimes think like mobs. Mob thinking can be short sighted and brutal. It falls for cat videos too.

[Outcast_Searcher]

I also recently watched it, and recommend it.

I both look forward to the upcoming AI revolution, and at the same time fear it. I don't fear the AI, more the people who will control it.

Thanks for the recommendation on the documentary. I'll be sure and check that out.

I was thinking that IF the AI objectively would measure the "right" things -- i.e. overall productivity, learning, progress, etc. -- then I think an AI could be a better boss than most human bosses. (I've had human bosses all over the map, and a great human boss is awesome, but of course terrible human bosses are hell. For me, the greatest motivator was what seemed to be "fair" reward for when I did good work over time. And good bosses would ignore the occasional screw-up (which with humans trying to get a lot done WILL happen) on performance reviews, if the overall impact of them was minor compared to your contribution. A good AI could potentially do that.

But your excellent comment about "the people who control it" is right on. If it's all about pinching pennies or prompt mindless paperwork, then those people would be just a form of "bad bosses", now armed with lots of objective data to, for example, nitpick away any raises or promotions.