At the mercy of AI: Your job, your health, your money

Topic

New Reply

If you’re concerned about personal privacy – and you should be – this article will take you back a step and look at a bigger picture. The focus on los
[See the full post at: At the mercy of AI: Your job, your health, your money]

Reply | Quote

Replies

That is a thorough and insightful overview of the problem.

Perhaps it is an article of faith, but I would contend that the problem is not technology per se but the use of half-baked assumptions and biases as the inputs, along with a general lack of responsibility for the results.

Credit card companies monitor the types of purchases you make with the card. Nothing stands in the way of their sharing that data with health insurers, potential employers, anyone with an asserted need to know.

In the railroad business I encountered an insurance industry practice of basing railroad liability insurance on payroll. Again and again, I complained about this to the agent. Apart from the obvious connection between employee lawsuits and insurance claims, there was little or no connection between payroll and liability. We were engaged in a large-scale effort to upgrade our track and other facilities, with significant benefits to safety, yet here we were being punished because it required us to add people. My complaints won us a temporary break on pricing, so long as there were no claims, but the basic method of setting these insurance rates continued. Garbage-in, garbage-out.

A few months ago I did some research on credit bureau reporting rules after a bill collector working for a medical billing contractor sent me a collection notice for a $7.36 balance-due bill which the billing contractor subsequently admitted in writing was never rendered. Credit bureaus are required by law to investigate when notified by the consumer that their information is erroneous, and they are required to follow up with the consumer confirming or denying the consumer’s claim. Apparently, however, most credit bureaus are unable or unwilling to investigate because the data they maintain are so sketchy that they cannot comply with the law.

So, in effect, these organizations are selling unsubstantiated opinion, fed to them by other parties, that if voiced by an individual might be cause for a defamation suit. They are not held responsible in any meaningful way–the penalties are peanuts.

This is one of the best reasons to try to protect oneself against online snooping.

Reply | Quote

Agreed. And one of the reasons why I’m very ambivalent about having ads.

People should be able to look at their snooping report, just like a credit report, and remove anything they don’t like.

Microsoft is missing a golden opportunity here – and it would start by MS documenting exactly what data it’s collecting, how to turn off the spigot, and how to review and modify (or at least delete) the results.

Ain’t gonna happen. But imagine if it did.

Reply | Quote

This is so depressing but predictable, human nature never gets better and greed will be the one thing that fuels this.

I’m one of those (to use Woody’s amusing analogy in another post) that’s “been around the fire hydrants” quite a few times. Back in the 80’s in the UK we had the ZX81 to cut our coding teeth on.

In comms, Viewdata then came along (Prestel) and of course Fidonet for which I was a NC for some while, long before the Internet came about.

It was real “steam computing” in those days measuring RAM in Kb but the main thing I remember is that we all had such optimism about where it was leading in the future.

Instead it has become misused simply for greed by those who already own more than they could spend in a lifetime and for social control.

I should have known better but I guess all old people say that when they look back.

I have rambled enough, but I will say this, age has caught up with me so I’m now approaching the end of my life and for that I am eternally grateful. At least I won’t have to see the worst of what will come, but for what it’s worth I am sorry for any part I may have played (hopefully little) in bringing this about by working in IT.

Reply | Quote

Dave Ramsey, a financial advisor with a national radio program in the US, sometimes speaks of how banks do mortgages – they look at your FICO score. There is no human being who actually takes a look at the mortgage applicant, to determine credit worthiness; it’s all in the FICO score. Ramsey brags about the fact that his FICO score is zero. But that hinders him from being approved for a mortgage, even though he is a multi-millionaire.

Think about the census done every 10 years in the US. They already do some doctoring of the numbers, based on the idea that some people will be missed in the count. AI could easily suggest a more “scientific” and “accurate” way to doctor the numbers.

Another thing which wasn’t mentioned: voting. Suppose there is an election, and it is “checked” with AI. The AI checking program calculates who SHOULD HAVE won the election, based on “scientific” factors. If candidate A won the election, but AI is sure that it should have been candidate B, will some judge believe that there was voter fraud solely based on what the AI calculation said, or will there have to be actual evidence?

I agree with all of his proposed solutions.

Reply | Quote

One reason that companies will resort to AI rather than their own judgment will be to avoid lawsuits. If they make the decision based on their own judgment, then they might be held responsible for the effect of their decision. However, if AI makes the decision based on its own “scientific” analysis, then the company won’t be held liable (they hope) because they didn’t make the judgment call.

This will motivate companies to not reveal how AI comes up with its decisions, because the revealing of that information might put liability onto someone. If it’s all kept “nameless” and “faceless”, then there’s no one to blame.

For this reason, I predict that AI will become an ever bigger part of our lives, unless there are laws which hold it in check.

Reply | Quote

Another basic problems is the lack of sound theories underlying the algorithms. Its just unbaked correlations and interpretations that are likely spurious and fishing expeditions that persist until some “pattern” is found by some machine that is convenient and can be used to exploit

Reply | Quote

It is said that sunshine is the best disinfectant, however MS evidently feels that disclosure does not serve its business interests at this time. Some people are OK with the snooping but often these same people do not realize that they do not possess the needed information to truly know whether they are OK with the data mining or not. There really is not much possibility of an “informed consent” without sufficient disclosure. I am not totally against targeted ads because there are certain applications that are too minor, although convenient, too expect people to underwrite a pay wall. However, I do not consider the Windows OS to be one of those applications and MS’s hybrid Windows business model may not be sustainable.

Reply | Quote

About the ads — I placed your site on the Trusted List for Ghostery and Abine Blur. Now I do see the ads, and since I’m in Windows 10 right now, I assume you will get the full benefits.

I think you have put enough thought and care into selecting how and who will provide the ads, that privacy or other dire consequences have been minimized so far.

The ads are unobtrusive and would probably many of them would interest a tech oriented readership. Good job so far.

As for those Credit information companies, don’t get me started. Equifax still thinks I am somebody else, even after years of submitting proof of my identity and addresses. The others are more up to date, especially Trans-Union, which does have current and accurate info on me.

It is true, according to CBS 60 Minutes, that at least Equifax, these companies never correct anything, even after being successfully sued. The awards, no matter how large, just don’t seem to happen often enough or have great enough dollar amounts to change industry practices.

I suspect any successful lawsuits against Google, Apple, Microsoft, Yahoo or anybody else would suffer similar fates. This is getting pretty ridiculous, and it does impact us every day.

We pay insurance rates, rent cars, get loans and mortgages, all based on credit reputation scoring. Even access to Google accounts may soon depend on a “reputation score” about which Google reveals little to the public as to how they arrive at their scoring results.

And the litany of grievous and not so grievous injuries and insults goes on and on. What Cortana “knows” and stores is totally opaque to Windows users. Even what Google “knows” and stores is not entirely available for us to view and delete.

This is an areas which will need to be watched closely for a long time to come.

Reply | Quote

At least in the US, vote adjustments of that sort would be rejected by any competent court of law. It’s unconstitutional.

Of greater concern is that in some States, voting machines are electronic and leave no paper trail. This makes a physical recount of the vote impossible in the event of a challenge. In Illinois, we have paper backups, mandated by our State laws. There needs to be a Federal mandate to have paper backups.

US Census adjustments actually improve the chances of getting correct funding distributions for cities where many folks aren’t located or fail to respond to census canvasing. So far, the algorithms have been favorable for social programs, not unfavorable.

As for FICO, it is based on reports from the consumer credit information companies. I addressed problems with these companies in another comment here.

Reply | Quote

The only solution is for all of us to utilize secret identities for ourselves, our computers and our internet service provider connections.

If defending “Truth, Justice and The American Way” (TM) in disguise is good enough for Superman, I see no reason why any governmental or corporate entity should object to universal adoption of his methods.

Reply | Quote

I am not a lawyer, but I believe that the methods in such cases would be subject to subpoena in Discovery or at the time of actual trial. Even the BI vs. Apple case was affected by the fact that if the case had gone to trial, Apple had the right to demand to know how the FBI tried (or succeeded) at breaking into the iPhone in question.

It isn’t possible in the American trial system to claim that the algorithms and methods used in AI are protected “Trade Secrets”. Not if the outcome of a trial may be affected by the reliability of the AI predictions or findings.

Reply | Quote

Knock, knock, knock.

Clark Kent?

I’m Mr. Mxyzptlk…

Reply | Quote

In my field it is common to use regression analysis to determine, for instance, what operating activity best correlates with (“predicts”) variable cost behavior.

The claims that are sometimes made for using FICO scores and the like for such things as insurance strike me as pure supposition, like the use of railroad payroll to predict liability exposure. I question whether anyone has done regression analysis to back up such supposition.

Reply | Quote

“US Census adjustments actually improve the chances of getting correct funding distributions for cities where many folks aren’t located or fail to respond to census canvasing. So far, the algorithms have been favorable for social programs, not unfavorable.”

I’d rather have accuracy than funding for social programs, if I had to choose between the two.

Reply | Quote

Well if these so-called genius types develop a computer with superhuman intelligence I want mine with rounded edges (sue me Tim), waterproof, fireproof (Samsung out), germ proof, virus and malware proof, no EULA or quality rollups (bye bye Microsoft), snoop and ad proof (Google take a hike), can not stalk me (Facebook, Linkedin etc. totally banned), omnipotent and with a huge red emergency power-off button.

Reply | Quote

@David F,

How wonderful that you were involved in all that! What a great career.

You did not know how certain things would develop, you had great intentions and optimism.

—
It’s so sad that many of us think (quite reasonably), as you have written,
“at least I won’t have to see the worst of what will come”.

No wonder many people in our countries are depressed and full of existential grief.

I’m probably at an age, middle-aged, where I will live long enough to see some of these things getting worse and worse, irretrievably so, and it frightens me.

Reply | Quote

On a lighter note, can someone tell me what the fire hydrants analogy means? What is the scene/action that is being alluded to? I honestly don’t get it.

Reply | Quote

Doing regression without a theory that explains the relationship is a fishing expedition that can and usually does yield spurious results that are taken as validation ***of what***?

Reply | Quote

Thank you for posting the link to Galen Gruman’s article.

I learned from it and am so glad that information like this is being shared by journalists and industry experts.

Reply | Quote

https://www.youtube.com/watch?v=41gOdlh5l-s

Reply | Quote

But why will it necessarily be worse? If our new electronic overlords are constructed properly, all sorts of possibilities open up.

I imagine many people lamented the passing of the family farm – or bronze tools – with the same trepidation.

Reply | Quote

In my experience, responsible analyses start with hypotheses and test them using regression, not the other way around.

Reply | Quote

What is needed is a rebuttal to the piece that is the starting point of this thread, explaining why such optimism is warranted.

Reply | Quote

I generally don’t allow google sites, so I tried to find out more about the Youtube link at Startpage by using their proxy view, but didn’t get very far (which I didn’t expect to!) They did have the video’s description of “Only in NYC a pittbull enjoying a nice cold water from a fire hydrant on a hot summer day.”

I am not sure why you made the hydrant comment as a reply to my comment on the post of “anon user”. My comment could have been seen as being a little ‘high-horse’, though I think it wasn’t too strongly-worded, given the several instances of chastising (of us here collectively) that had come from his/her side before I wrote what I wrote. I just wanted to understand why it was portrayed as something that someone who’d been hanging around hydrants would write, but I’m okay with not knowing, it was meant in a humorous way, I’m sure.

—
Anyway, I feel a slight sense of foreboding tonight, not sure if it’s the dramatic change in the weather/temperature here.

I want to go vote (early voting) to avoid the crowds on election day, but there’s been some violence at a political site in my town (it’s not the town on the east coast that made the national news last week) and I also have a sense of foreboding about possible violence to come in our town, perhaps at our city hall (my voting site). There’s enough violence as it is, so many homicides and drug deaths here this past week. Hearing the ambulance sirens on nearby main streets gets one down. My heart hurts from the depths to which our national discourse, media, entertainment industry, politics, educational system, you name it, have sunk.

Maybe I’ll do my own personal Brexit, a Poohxit, where I’ll hole up in my tree with a few pots of honey and ride out the next few decades.

Oh would that I were there, tramping around the Weald of East Sussex.

https://en.wikipedia.org/wiki/Ashdown_Forest#Winnie-the-Pooh

https://en.wikipedia.org/wiki/Wealdway#/media/File:Kent_Long_Distance_Footpaths.png

complete with anonymizing “Mr. Sanders” nameplate for privacy!
https://en.wikipedia.org/wiki/Winnie-the-Pooh#/media/File:Pooh_Shepard1928.jpg

Honestly, I’m not a Pooh nut, I just picked this username out of the air and it’s the only place I use it.

I used to be called D. here and then D.D. although there isn’t anything “D” about my name, or anything else
(now it’s more become more like
“Carry On Windows”, or Benny Hill, oh dear).

[“The Carry On franchise primarily consists of a sequence of 31 low-budget British comedy motion pictures…. The films’ humour was in the British comic tradition of the music hall and bawdy seaside postcards.”
https://en.wikipedia.org/wiki/Carry_On_(franchise)%5D

Reply | Quote

Just look at history, humans repeat the same mistakes again and again and we are seeing it now.

The problem is not with the technology, the problem is the use it will be put to and if you really believe that large multinational Corporations are going to suddenly become benign overnight then I think you’re just living in hope not the real world.

Looking back there used to at least be a sense of corporate decency in the UK (I can’t speak for the US) which seems to have completely disappeared in the race to make a fast buck at any cost.

Humans will create AI in it’s own image greed and all, as the whole push is by those seeking to extract as much cash as they can from us, who are viewed as the product. If they did not think they could turn a profit they simply would not bother in the first place. The purpose a business is to make a profit and that is its only purpose.

The major problem now is having a few mega-corporations who now view the customer as a product to be traded, so no I do not see a reason to be optimistic as we are just repeating the mistakes of history.

Altruism does not turn a profit.

@poohsticks

Thanks you for your kind words.

Indeed in many ways it was a wonderful period to live through and to see the changes that came about.

Many may think of the valve evolving into the chip as the driver but there was much more particularly in the fields of materials technology which (like glass in historic times) really changed the world

Many things we take for granted now were just “Science Fiction” when I was a child, so it has been a remarkable and sometimes privileged journey through life.

Reply | Quote

It was meant in a humorous way.

Reply | Quote

LOL, when Woody came out with it I just assumed it was an equivalent of a “few times round the block”

Reply | Quote