borg.org

Um, …Yes.

People think the questions of AI ethics arise from dealing with the consequences of what Turing and all those cats invented: Computers. And they are right, to a point. But more it is about dealing with the consequences of what the Sumerians invented: Bureaucracy.

Teachers infer the emotional state of students all the time, so what if a computer program does it, what is the problem? If it is okay for the human to do it, then it is okay for the computer to do it, too—as long as the computer does it competently, as long as the computer does it fairly. Right?

Wrong.

When a teacher looks a student in the eye and infers his/er emotional state that is an invasion of the student’s privacy. And, as an isolated act, it would be wrong. But it is not an isolated act.

• The student understands s/he is in a school, has frequently chosen to be there.
• The teacher is looking into the student’s eyes (or otherwise observing, maybe asking questions of the student), and the student can see that.
• When the teacher is looking into one student’s eyes s/he can’t also drill into another student’s eyes.
• And the student can look back into the teacher’s eyes, and the teacher can see that.
• The student gets to infer, too.

All this is part of a larger human relationship. And that is valuable. (No, not all human relationships are good, but they are pretty much all we got.)

Estimating a student’s emotions is just a small part of what is going on between a teacher and student. In isolation, staring at someone is an aggressive act, one where the “student” might be justified to hit back, but more wisely would turn away from this crazy person, cross the street, and put some distance between them.

Just because a human does thing X, even as a key part of some wonderful accomplishment, doesn’t mean thing X is itself somehow good or even remotely acceptable in other circumstances.

The fundamental problem with Intel’s innovation has little to do with whether they did a good job of implementing this isolated skill or not, it is more the fact they implemented it as an isolated feature that then can be deployed in myriad ways. Ways the subject (“student”) maybe has no knowledge of, ways that can drive decisions about the subject that s/he has no ability to influence nor appeal.

When used completely as intended, as part of online “Zoom” learning, this innovation might not be all bad, but the student is no longer in a student/teacher relationship of the sort we think we understand. No, this rapidly evolving online tool that Intel apparently wants a piece of, is redefining what is going on. And God knows what “data driven” innovations will be added next, particularly in the hands of for-profit institutions.

The real problem here is what horrors are possible in the maw of faceless and unanswerable bureaucracy. The fact that the bureaucracy has a shiny new toy is important, but don’t so focus on the toy as to ignore who is using it, and how it will be used as part of a much larger system.

I have seen the movie Schindler’s List only once, it was in first run, a zillion years ago. I do remember it was in black and white, and I happened to see it on a glorious big screen. But I don’t remember a lot of other details. Yes, there were nasty Nazis, terrible dilemmas, sadness and heroism, but it’s all kind of blurry. Except one chilling shot etched in my mind: a typewriter, typing up names. The film hit me between the eyes with that.

Bureaucracies have always kept lists of names, the difference is the typewriter made it more efficient. Everytime bureaucracy is handed a more efficient tool, we are at risk of bureaucracy doing what it does at a more industrial scale. As impressive and powerful as the typewriter is, seeing it in the movie also reminded me that IBM sold more powerful information processing equipment to the Nazis. To help them run their bureaucracies more efficiently.

Some of the the new computer programs we like to call “AI” are really impressive information processing tools. They are tools that allow bureaucracies to not just possess and sift through vast amounts of data, but tools that allow bureaucracies to make subtle decisions about that data without bothering with the pesky bureaucrat step anymore. No human bureaucrat means breaucracies can now make complex decisions at scale, at great speed, and therefore make many more decisions than ever before.

Sure, individuals can deploy this technology in bad ways, but so can bad teachers can be bad in completely old fashioned ways, individuals don’t need such a fancy tool to do individual harm. But bureaucracies need to operate at scale.

“AI” ethics isn’t about the “AI”, it is mostly about bureaucracies—government and commercial, public and secret—and ethics in how and what they deploy. Bureaucracies have never been good at ethics.

But right now they are all transfixed by this shiny new toy, looking for what data they already have sitting around, what new data they might collect, and how they could plug it all together, to do things they could never do before.

Read Kafka’s The Castle, then let your imagination roam…

-kb

©2022 Kent Borg

P.S. Comments are broken and have been for sometime. Sorry.

Today, 19 May 2020, is Oliver Heaviside’ 170th birthday.

Hurray!

Who?

Okay this post will be a bit technical…but only a bit.

One of the giants of science is James Clerk Maxwell, he figured out the science behind how radio waves work. Pretty important stuff. He is immortalized in the famous four Maxwell Equations. Maybe you have head of them, lots of science-y types have heard of them, but not that many understand them, so don’t feel too left out, this is rarefied territory. (I’m still learning this stuff, but I’m not there yet.) But it seems they are important to the experts that make modern life work.

And Maxwell did them!

Except he didn’t. Maxwell did eight equations, not four. Oliver Heaviside is the one who simplified the eight into four.

People will write otherwise, but it seems to me that it was Heaviside, not Maxwell, who created the whole field of electrical engineering.

In Heaviside’s day the hot new technology was the telegraph. You know, Morse code, dits and dahs. It was slow and cumbersome and expensive, but it could communicate over long distances very rapidly when compared with a horse or ship or even a racing locomotive (that means a train). The slight detail is that the longer the wires got the mushier the signal got. People would try to make up for it by turning up the voltage, and they had other tricks they tried, but it was a seat-of-the-pants, rule-of-thumb world, and don’t ask too many questions because even the most talented “electricians” didn’t really know why one trick might work and why a different did not. The first trans-Atlantic cable was not only very slow but burned out after a very short amount of use. (They turned up the voltage quite a bit.)

Heaviside worked for a telegraph company, and he wanted to figure out this stuff worked. Precisely how it worked, as in quantifying things. This rubbed some in the industry the wrong way, there was a lot of opposition of quantifying these things. And he did. Both rub people the wrong way and figure out these things. He figured out how to make telegraph wires operate at much faster speeds for much greater distances, without burning them out. His same principles were applied to voice telephone calls, for they had the same problem of the signal getting mushy if the lines got too long, and he explained how to fix that, too.

His solution was to set up what is now called a balanced transmission line. Back when TV antennas were put on roofs the first kind of cable for connecting it to the TV was a balanced transmission line, 300-ohm twin lead, to be precise, and it was a flat cable, made of plastic with a wire running down each edge. Don’t tape it directly to your metal antenna mast, it doesn’t like that, but if suspended away from metal, it is very efficient at getting a very weak signal down to the TV without picking up interference along they way. Heaviside invented the balanced transmission line. And it is useful for a lot more than ancient TV antennas. It you haven’t heard of balanced transmission line, but have heard of coax cable (it is perfectly okay to tape coax to the metal TV antenna mast–coax is not as efficient as twin lead, but easier to string), well Heaviside invented coaxial cable, too.

About the time telegraph and telephone were still pretty new there was another hot technology: radio. Heaviside was paying attention there, too. At this point sensible people knew the world was not flat but is round (spherical, to be pedantic). And people also knew that radio waves travel in straight lines, so radio wouldn’t be useful for long distance communications, right? Wrong. For some frequencies, under the right conditions, radio waves will bounce off the ionosphere and can travel great distances. Heaviside figured this out. He figured out that this should work before people found out that it did work. In fact the layer of the atmosphere (the ionosphere) that does this was originally called the Heaviside Layer.

Back to the Maxwell Equations. The way that Heaviside did all of this was by taking the science and figuring out how to apply it in a precise way to make an engineering discipline: he created electrical engineering. Including inventing new ways of doing the math.

In the early days of Bell Labs they were working magic by taking Heaviside’s work and applying it in a practical way. Some Bell engineers were so impressed with Heaviside’s work, and so indebted to him, that they tried to send him money, but he said no.

At this point Heaviside was old and not rich, yet he said no. Oliver Heaviside could be difficult. Various folk tried to help him and to the extent it looked like help to him, he said no, even though he needed it. And part of why the Bell Labs engineers were having such a hay day on his work is he that, though he might have been brilliant, he didn’t stop to try to make his work easy to understand, it took awhile for his work to have full effect.

Back to the my claim he should be the patron saint of ham radio: He made practical much of what the field is built on, he was a hands-on man. And he died of complications from falling from a ladder.

-kb, AC1HJ

©2020 Kent Borg

P.S. Comments are broken and have been for sometime. Sorry.

As we try to tighten up our computer systems, in 2018, phishing feels like one of the most dangerous things. Sure, getting someone to open a dangerous attachment that exploits a PDF bug (is there an infinite supply?) is a problem, but let’s imagine users running on such tight systems that dangerous attachments are no longer a problem. Phishing won’t be over.

People will still be fooled by crooks, and if a crook walks up to you in nose glasses and politely asks for the keys to your car you might think something is funny. You do expect give your car keys to strangers, but you expect these people to be, say, the mechanic at the car repair, or the parking valet at the restaurant. The point is if you initiate the transaction (you go to get your car fixed, you go out to eat), the transaction is much safer. In contrast, if a stranger approaches you and volunteers to be a mechanic or valet, you are less likely to fall for it.

We should apply that to computer credentials. If I decide to go to website tuklever.com, it is reasonable for me to then type my TuKlever login credentials. But if some website claiming to be tuklever.com approaches me (via e-mail), why should I hand it my password (aka my “car keys”)? I shouldn’t.

A powerful way to avoid a lot of phishing attempts is:

Never (never ever) type credentials because someone else supplied you with a link, possibly in an e-mail.

Instead, if the link looks good, login to that website manually (type a trusted URL by hand, use a bookmark you typed by hand). Now log in. Now try that link in the e-mail–and if you are asked for another password, don’t do it.

This logic applies in other circumstances, too. Get a call from our credit card company about a suspicious transaction?, and that professional sounding voice asks for verification information from you? Say no, ask what it is about, say you will call back. Call what number? The number the voice on the phone gives you? No! Call the number on the back of your credit card. Same idea as the website example.

Back to my headline: We train people to do the dangerous thing.

• Employees frequently get real e-mail from, say, HR, that includes links to, say, the new payroll system, and click and type in sensitive information.
• For years now American Express has been sending me e-mails that include links to click on and an invitation fo type a password.
• Other credit card companies–fraud departments even–have called me and expected me to give them identifying information.

In each of these cases we are conditioning people to do the dangerous thing. In each of these cases the safe thing to do and the normal thing to do are different. Do we really expect people to rock-the-boat, and refuse to log into the new payroll system, and not get paid?

No. We expect people to be phished, and we are training them for that.

I realize there is a heresy in what I am saying. I am implying that user behavior matters, that how we condition users matters. I am a hairsbreadth from suggesting that user education is a good thing! Horrors, the first step down the slippery slope of blaming the user for bad system design. Next thing you know I’ll make a snide remark about some celebrity caught on live camera entering the PIN 000000.

-kb

©2018 Kent Borg

Linux has won. It is taking over everything, from tiny devices to the biggest super-computers. Apple’s operating systems are all pretty much on the same model, and Microsoft always seems to be trotting along in roughly this direction, too.

The idea is pretty cool: Give each program a uniform view of the machine, keep them from interfering with each other. Not only can each program mostly pretend it owns the entire machine, the model is good enough to be extended to multiple users, all running on the same machine.

Yes, there will be resource limits with all this sharing going on, but that is a necessary limitation, the larger sharing model is great.

So why are we so unhappy with it? Why do we have this big virtualization fad? The operating system was supposed to let multiple users share the same physical machine, why an extra layer of multiple operating systems sharing the same hardware? If these multiple operating systems were different kinds of operating systems (needed to be compatible with different kinds of programs) that would make sense, but mostly we run multiple virtual copies of the same operating system. Frequently the same version of the same operating system. The popularity of hypervisors for providing multiple uniform views of the hardware, keeping them from interfering with each other, seems a big indictment of what the OS was supposed to do. Something is wrong with the API offered by the OS if we prefer the API offered by BIOS. Something is wrong.

And inside the OS, different programs were supposed to do the different things. So why are we now inventing enormous container facilities like Docker and Kubernetes for supplying the features we want? Isn’t that what the OS was supposed to orchestrate?

I don’t see much questioning of the role of the OS, but I see an awful lot of ad hoc reinventing of OS-like services.

Part of this is clearly a limitation of the OS model: Individual programs are isolated from each other, but it seems not isolated enough, we want more isolation, so we fire up new OS instances. Also, individual programs have complicated and conflicting dependencies to shared libraries that the old OS model isn’t good at mediating. Finally, individual programs are not where the action is, we run different programs in concert with both dependency confusions between them, and contradicting desires to be isolated from other programs (so they don’t interfere) but not isolated from other programs (so they can cooperate). It seems these are all issues the OS should handle, and it doesn’t, that’s why we have so many VMs, and these container facilities.

Recently I ran across the various name-spaces that the Linux kernel offers. (Linus is very pedantic that the kernel just be the kernel, but that doesn’t mean it isn’t still freaking gigantic and bursting with features.) These name spaces provide a lot of granularity for controlling what is isolated and what is shared between different programs. It seems they make it possible to completely isolate software, as if you were running completely different operating systems. I say “it seems” because I don’t know that I am right, I don’t know that these different name spaces cover all the bases. And, even if I knew they did cover all the bases, how would anyone ever trust that they did it in a bugfree way? How would anyone ever know that there isn’t some unintended leaking between spaces, security holes hidden in the confusion.

I think this gets to the point: The confusion. The old OS model was simple, that was a virtue. The model implicit in Linux name spaces is so complicated that I almost don’t want to call it a model: if almost no one can understand the implications of all those features, can it be called a “model”? Does it instead become an “artifact”? Something to be studied, as opposed to a model, something clear enough to be understood?

Maybe I am just being overwhelmed and demonstrating my ignorance. But something tells me that the simplicity of hypervisors, presenting a near bare-metal model, isn’t about to lose its appeal as everybody starts to grok Linux name spaces.

I think we are choking on unmanaged complexity, that we are building systems that are more complicated than we know, that not only are they riddled with conventional bugs, but attackers are waltzing though our systems via the security holes made possible by that complexity. But that’s another topic.

My conclusion here is we have run out the old OS model to the point of absurdity, that we need to rethink what abstractions an OS should offer. The old OS model was both powerful and simple, but look at the layers of baroque filigree we are accumulating, it is time to revisit our assumptions about what an OS is.

-kb

©2016 Kent Borg