Thursday, March 15, 2018

Makin' Memories III

So what's the point of all this noodling and rambling on about the philosophy of photography and its fascinating relationship with <yawn> contemporary social conceptions of <yawn> portraiture?

Well first of all, welcome to my blog, you should be used to it by now. But also there might even be a useful nugget in here.

Business people, the good ones, know that you need to know exactly what the hell it is you do as a business. It's not delivering shareholder value, it's not building the finest automobiles. What is is that you actually do? What problem are you solving, how are you doing it, and for whom do you do it? This informs almost every aspect of how you should run your business.

You can also retask this sort of exercise to a hobby, or a passion. What, exactly, are you trying to do here?

Let us suppose that you are a product photographer, or would like to be one. What do you do?

I take pictures of products.

Well, sure, but drill down. Anyone can take a photograph of a widget or a candy bar. I can do it with my phone.

I take pictures of products that make the product look good?

Better. Look good to whom, and how? What does "look good" mean?

I take pictures of products that reflect the client's brand identity, and which make the product look appealing to the client's potential customers.

Now we're gettin' someplace. This is a statement that not only reflects what problems you're solving for your customers, but also suggests some things you might do to up your game, to do a better job at what you do. For instance, you could go study up on corporate branding and identity. If you read a book or two on Branding, suddenly you're talking the same language as your clients in a new, important, dimension. You can make suggestions that are aligned with what you actually do.

Let's get back to these businesses that take photographs of people performing. Engagement sessions, wedding photographers, Senior Sessions, that kind of thing. Once you've identified your job as:

I take pictures of people performing improvised scenes based on their real lives.

again, you're in a position to up your game. Sure, you could spend a couple grand on a Sigma Art lens for even more creamy bokeh. But let's say you've only got $180 this quarter. Let me make a suggestion: Hop on MasterClass and buy yourself a 1 year subscription. Then take the online course from Ron Howard on Directing. Yeah, Ron Howard, the guy that made basically every movie. One hundred and eighty bucks.

I have no affiliation with Master Class, they've never heard of me, and, god-willing, never will.

But unless that course consists of Ron Howard silently staring at the camera while picking his nose, there is basically no way you won't get $180 worth of value out of the class. Or, if you don't have $180, see if your local library has a couple books on film-making or directing.

If you're basically a photographic taxidermist, like so very very many of these store-front portraiture operations from the 1980s (or run by people who learned at the feet of a 1980s era taxidermist), there might be no hope for you. But then again, Ron Howard's a damned engaging dude. Worst case, it's probably $180 worth of sheer entertainment.

Wednesday, March 14, 2018

Makin' Memories II

Carrying on. I think what I'm really picking at here is what people are and have looked for in things that resemble formal portraiture.

From time immemorial, people have commissioned painted portraits, and they still do. It was, and remains, a act that carries a couple of meanings. On the one hand, it's a projection of wealth and power, and on the other hand it preserves a likeness of the subject. The subject demonstrates wealth, and has a shot at a kind of immortality, if you will. Poor people never had a likeness made of themselves, of any sort.

Fast forward to the 19th century. There's an explosion of portraiture with the advent of the photograph. Even though the various processes were tedious and in modern terms expensive, the ability to have a likeness made became accessible to the middle class, and then even the relatively poor. While it may have maintained a thin veneer of social status "Oh, what a lovely quarter-plate tintype. You know, I had a half-plate done." it mainly shed that aspect of wealth projection.

Of course, we still have even today the social status, but it's landed on a handful of practitioners. Getting shot by Snowdon or Leibovitz would certainly be a social feather in the cap, as well as an obvious "tell" of wealth and power. As a general phenomenon, though, general portraiture shed that in the Victorian era.

The purpose in the Victorian era seems largely to be simply creating and possessing a likeness of the subject. As commenter remarked on the previous, on the memento mori photos of this era in which the dead are photographed with the living. If the point here is not a shot at immortality in some sense, I cannot imagine what is. The point here is to preserve a likeness, to remember the subject's appearance. Sally Mann had some interesting remarks on memory and photographs which might be a sort of salient side note here, the bit where photographs destroy memory.

This is, as near as I can tell, roughly the state of affairs through the 20th century. People go to "get their pictures done" with gradually decreasing enthusiasm for the next 100 years or so. More and more people are taking snapshots, but the results can be a bit unsatisfactory, there are no guarantees because you're an amateur using film and, let's be honest, not everyone is that in to it. Still, vernacular photography arises, and people get used to the idea that "lots of pictures of the kids" isn't a weird thing, or reserved to the wealthy, or any of that.

Enter digital. Everyone's a photographer now. There isn't any cachet whatsoever left having a photograph made of you, it happens more or less constantly. There's no rational reason to look for more likenesses of anyone. If we want to remember what so-and-so looks, or looked, like, we can just hop on Facebook or Instagram and see 100s of pictures.

The forces that drove the Medici to have their paintings made have vanished completely, except, one supposes, for inertia.

At the present time people taking pictures for money are all having a rough time, you have to find some angle to make it work. The standard line appears to be something about "quality" with some vague handwaving about equipment, experience, and passion. That doesn't actually work very well, but there are some things that have some traction.

By tying this portraiture business to specific times of life, to events, you can brings in some customers. "Seniors! Get your pictures done! Preserve the memory!" "Engagement Sessions" and so on. Almost any sort of specific reason, beyond "get your picture taken", will pull in some customers.

That said, standard portraiture is a thing you can still buy.

Let's interview me, I'm right here after all.

You could pay me to go down to one of the local studios, but my hourly rate is still north of a couple hundred bucks.

But I'd sit for Kirk Tuck in an instant, and if it wasn't tremendously inconvenient to get to his studio, I'd probably have pestered him to shoot me already.

The difference, for me, is that the shoot at the local studio is likely to be an itchy, boring, experience in which I stand in front of a mottled colored backdrop from the last century while some perfectly nice but dull fellow fusses with lights and waits for me to put on my Camera Face so he can shoot. The experience with Tuck would involve tea, conversation, and mostly no mottled backdrops. It would be interesting, because Tuck is an interesting guy, and because he knows that the pictures aren't going to be any good unless he makes it interesting.

The result would be in some sense just another likeness of me, of which the world already has an excess. It would be, if we had the wind at our backs, also insightful and in some way revealing, it would be in some sense beautiful, and the memory of making it would be of an enjoyable collaboration. Not itchy at all.

Also, I am a weirdo, so this sort of thing appeals to me even though as an introvert I find it exhausting.

As for events, I have one (1) photograph from my actual wedding, which I like a lot. It is an artifice of sorts, in that we all piled together into a sort of Last Supper arrangement for a Group Photo. It is also a real moment, because there we were at the wedding, just married. The photograph serves to evoke the memory of that terrifying day when my life changed and became what it is now. I'm glad I have that picture, but I don't need any more. (more were taken, that's the one we chose to print and hang)

So there has been, socially, a pivot. We're no longer getting from our portraits anything of what the Medicis' got. What we're getting now is much more tied to events, to times and moments in life. To be fair, there is a genre of painting which we might loosely characterize as "The Duchess McStuffing On The Occasion Of" which was likewise tied to events, way back in the day, but I dare to speculate that it wasn't the Main Thing driving painted portraits.

Now it is. There's an entire genre of retail photography with all the itchy posing weirdness of a painting (albeit on a somewhat shorter time scale) that is intended to evoke specific memories of a specific time -- which often isn't the time the picture was taken.

At this point we have this curious phenomenon where, a great deal of the time, the role of the professional photographer is to direct the subjects through a sort of play, which is then recorded by the camera. It's a sort of documentary, but with a great deal of re-enactment. It feels natural to people, possibly because we're steeped in it. This is after all also the essence of a great deal of commercial photography.

We've arrived at a curious juncture in which the camera, whose claim to fame is that it records reality and truth, is called upon to record instead artifice and re-enactment. While it records, naturally, a likeness, the point of the exercise is to record the artifice.

Personally, I don't much like it, I think the seams are altogether too often much too visible. I believe that the reasons the seams are too often visible is because the entire enterprise flies in the face of photography, philosophically, and that therefore to pull it off well you need a great deal of skill in areas that all too often the photographers don't even recognize as areas of expertise.

While satisfying to the customers, it looks fake, because it is fake and because the photographer doesn't have the directorial skill to extract something else.

To be fair, extracting naturalistic performances from people is widely recognized as monumentally difficult. Ask your theater nerd friends.

Tuesday, March 13, 2018

Makin' Memories

People say that photographs preserve memories. It's a trite marketing pitch. It reminds me of the curious little exchanges that occur when someone dies, let's say, skydiving. Someone will say "at least he died doing something he loved" and then some wag, usually me, will express surprise that he loved slamming into the ground at 120 miles per hour.

When you go get your photograph taken, you're not actually there to preserve a memory. The memory would be of being photographed, after all which is generally not what is depicted, and generally not the memory we're looking for. At least not with formal portraiture.

Here's a picture from John, who lives in Sooke, BC (which isn't quite as remote as it sounds, it's just around the corner from the provincial capital, which I must admit, is kind of remote).

Now, John is a professional with 30 years of experience, and he helps out on one of the more dysfunctional internet forums I know of, but he can stick lights into places, and achieve accurate focus. This picture makes my teeth itch. I hate this sort of crap so very very hard. Not only it is wildly dated and weirdly framed, the guy in the picture is just doing his Camera Face, with his mouth hanging dopily open. There's no precious memory here, there's just a horrible picture, and probably a sort of tedious afternoon. I dare say the customer is perfectly happy, basically because it's "clear" and it looks like him.

The picture surely pleases him, he's got all his stuff in the frame there and whatnot. The specific memory it evokes is tedious, but the general associations are probably pretty positive.

Here's another photo, probably some engagement session:

Again, we have the artifice of a "photo session" in play here. Most likely the couple has dressed up specially for this photo shoot, most likely the photographer is directing/posing (you can take courses on how to pose Happy Couples so that they Look Happy). Still, it is at least taken at that time of life, when the whole relationship genuinely does have that flavor, the freshness and excitement of the recent engagement. While the couple is in truth playacting for the camera, they're at least playacting a scene they believe in.

Again, the customers are without a doubt delighted. They've got, probably, exactly the pictures they had in mind.

They went out that day to help make these pictures, they worked at it, and here they are. The salient memory is of work, of standing in a field trying to hear the photographer's instructions while also trying to look beautiful and feeling a little weird and itchy wearing these clothes out in the middle of a field. The work, however, was successful. Why in God's name you'd want these pictures is a bit of a mystery, but they are a thing people seem to genuinely want.

Anyways, it is much the same deal. The specific memories evoked are likely tedious and not very comfortable, but the general sensation, the idea behind the pictures, is pleasing. At least for the moment.

I do these things with kids from time to time, with some moderate success, and never for money. So, I, too, am guilty here. The pictures made are quite nice, and record the appearance of the child, and even a flash of personality on good days. The process is awkward, not fully pleasant. The specific memories associated with the pictures, probably not wonderful, but the general era evoked is, well, it is whatever it is, no?

Twenty years from now, I dare say the memories of the "photo session" will be vague, but perhaps the memories of that time in the relevant lives will pop back in to focus when the subjects gaze upon the pictures? That's certainly how it works for me. Pictures of me as a kid evoke no recollection at all of being photographed, but to be fair often very little to no recollection of that time of my life either. A few later pictures evoke the time, without evoking the photographer.

Anything resembling a formal portrait, though, evokes memories of itchiness, discomfort, a generally weird sensation. We didn't do much formal portraiture.

What about a successful proper portrait session, as opposed to some low rent retail shop?

Ideally, this is going to land someplace a little further in the territory of an actual pleasant memory. Here's a guy I know, a stained glass artist I did a little fluffy interview with for our neighborhood "newspaper":

What's Erin's memory of this experience? I don't know, but I will hazard that it was a little sensation of being "on the spot" a little "arg, there's a guy taking pictures" but also a memory of a mildly pleasant conversation about his work. Erin was working as I shot, I have a bunch of pictures of his hands, his work, and so on. This is just a moment when he felt the camera come up, and so "posed" for just an instant, gave me a bit of a camera-ready-grin, and then went back to work.

We didn't go to a place where photography is done, I went to his place where stained glass is done. And then I made him a bit uncomfortable, certainly. But surely not as awkward as the fellow in the uniform up at the top. And it certainly wasn't as big or intrusive a production. The result, I am biased enough to say, is about a billion times better even though the lighting is a bit iffy and, to be honest, the focus is not in the right spot at all. Kirk Tuck would have hit the technicals much better, and most likely done at least as well on the "moment of genuine personality."

Next up we have candids, or near-candids. These girls are just hanging about enjoying a sunny day. Only one of them even sees me, she said to her friends "he's taking a picture!" immediately after this, to get them to look up and smile. I took that picture too, but it wasn't as good.

The memory here would probably be of waiting around with girlfriends on a glorious sunny day for, well, whatever they were waiting for. I hope they were having a good time, and I like to imagine that the guy with the camera was a bare ripple on their day.

Anyways, the point is, assuming I have one, that the "memories" associated with a photo are kind of all over the place. There's memories associated with the actual taking of the picture, but also then memories of that era of life, of whatever the picture was intended to capture and preserve. It's complicated.

Wednesday, March 7, 2018

Olde v. Newe

I read, when I read, quite a lot of photographic history. I get exposed to a lot of old photographs, therefore.

Man, portraits then look nothing like portraits now. I think we've definitely lost something along the way. Not only do I like the vintage style pictures just plain more (that is, I think they're flat out better) we've lost a lot of vocabulary along the way.

Now the only question is what appears behind the sitter. Is it out of focus? And by how much?

There's no question of whether the sitter's face should be sharp. Of course it is. You have to nail the focus on the eyes or your picture IS A FAILURE. You can't bury stuff in shadows, especially not parts of the subject! You have to fix the skin! You MUST! It's no GOOD if you don't fix the skin!

To that end, I have been practicing, off and on, producing things that look vintage. Yeah, yeah, it's a bit twee, a bit overdone. And yet, it's not done at all, except in stupid Olde Weste Portraite Studioes where everyone has hastily donned ill-fitting period garb and is mugging like an idiot.

My four year old (mugging). Olde and Newe.

Olde obviously has the advantage that you can conceal a host of sins. But I'll be damned if I don't just like it better, too.

The Bolt is not The Ship

Bolts are cool. There's a surprising amount of engineering going on there, and a lot of stuff to talk about when you've got a bolt and an application for it. There's tensile strength (how hard it it to pull apart lengthwise), shear strength (how hard it is to tear sideways), stuff about threads. When you tighten down the nut to a specific torque, you're applying tension to the shaft of the bolt and "using up" some of the available tensile strength, but there are reasons for why you might do that, and so on. There are probably entire textbooks on the subject.

The stuff we build, like ships, uses a lot of bolts. Naturally, when we're designing a ship, we spend a certain amount of time thinking about bolts and bolt engineering. Actually, a surprising amount of time. We don't want the damned thing to fall apart, after all. That's super awkward.

But the ship is not the bolt, nor vice versa. When we're designing a ship, we're doing Naval Architecture, which is a complete discipline. It includes much, only a piece of it is strength of materials, only a little bit of it is about bolts. The Naval Architecture piece allows us to calculate the way our bolts will be loaded, which enables is to do the right bolt engineering.

Again: the Naval Architecture is what we're doing. The bolt stuff arises as a consequence of the ship's demands. Occasionally, we might find that our ship doesn't even use bolts. Everything is glued or welded.

And so it is with the photograph. There's a lot we can say about the photograph, about groups of photographs, about how these things work. While these notions are often necessary to understand how, say Instagram works (as a social construct, how it works "in total" as it were), or how Law Enforcement Surveillance works, they are not sufficient.

In order to make sense of how law enforcement's gradually increasing photographic surveillance works, you need to know quite a bit about law enforcement, social systems, bureacracies, and also a bit about photography. The last one on the list is not going to magically imbue you with enough understanding of the previous. Indeed, we will only really deploy our understanding of photography when it becomes salient as we work out what the hell is up with the police. We might find, sometimes, that while visuals are in play, they're not even particularly relevant. Our dissertation on "index and representation" never even gets pulled out of its drawer (to our vast disappointment).

I am, naturally, as guilty of this as anyone else. But I try to be conscious of it, and I try to do Naval Architecture as appropriate, rather than hoping that my understanding of bolts will carry me safely through to the complete ship design. Probably not 100% though. Dang it.

Friday, March 2, 2018

Reading The Pictures on Trump and the Media

More or less intelligent people are passing this... object... around. In it, Michael Shaw of Reading the Pictures does, well, he does something or other. It is worth noting that a great deal of what he's doing in this piece of quoting himself. While he mentions at the beginning that

Note: Many of the examples were analyzed on the Twitter account for Reading the Pictures, a non-profit media and visual literacy site I publish.

You actually have to be paying attention to notice that the bulk of the article reads like "here's a tweet, look what it says. SEE?!" where the cited tweet is actually just the author all over again. This is a disingenuous technique, at best. Rather than quoting your own pen name as if it were another person, just re-state what you said before, bruh.

Michael Shaw, as usual, appears to have no thesis, or indeed any organized idea of what his point is. He's just blathering TRUMP SUX which, ok, Trump does suck. But we knew that.

This appears to me to be a set of cherry picked photos that Shaw feels create some impression of our sitting president.

Is Shaw saying this is a true portrait of the man? Not as far as I can tell.

Is Shaw saying that the media is constructing this portrait of a man? Maybe. Maybe not. If he has a thesis at all, this is probably it. He seems to be promoting this as "what the media is portraying" whatever that means (obviously there are large and powerful elements of The Media that are doing no such thing).

If this portrait of the man be false, then where arises the falsehood? Is someone manufacturing it, and if so, for what reason? Is the media building this image of Trump, or are they merely reflecting some current widely held notion of Trumpness? Is there any analysis in here at all?

In reality, of course, Shaw is simply constructing an image of Trump out of selected pictures, and then quoting himself as evidence that this is how Trump is being portrayed. It would be silly to ask any of the questions above, because they're all irrelevant. This is Shaw engaged in pure intellectual masturbation, and publishing it in an allegedly real journal.

I did not know that Columbia Journalism Review was a proper venue for reviewing ones own tweets. I wonder if they'll give me a column?

Monday, February 26, 2018

Fetishizing the Digital II

In the light of what I intend to write here, you could argue that I am outright contradicting some parts of what I said in the previous remarks. You could make that argument. I'm not gonna make it for you, though!

So there's another aspect to "the digital" that gets dragged out which I am going to pretend I was saving for a later post, rather than outright forgot about. That aspect is that the digital photograph is machine readable. Which isn't in and of itself a huge paradigm shift. You could probably have built something or another in Victorian era that would "read" a picture in some whacky sense or another, although I dare say nobody did.

Be that as it may, whether by a series of tiny steps or in one giant and shocking step, we do find ourselves increasingly in a world in which we, and our things, are photographed automatically and the pictures fed to a machine. Your automobile license plate might be photographed at toll points on roads or bridges, or by police officers, the numbers on the plate to be fed directly to a computer for billing or various law-enforcement searches. We either are at or on the cusp of our faces being photographed and fed into facial recognition systems constantly and for various purposes.

Something I probably won't talk about now is that much of what we shoot ourselves is also fed directly into the machine, but let's stick to the pictures machines take for the nonce.

At this point we've arrived at the usual point, in which boffins are inventing new ways to ingest photographs and process them for no reason except that boffins gotta boffin and "I have a cool idea!" The shady buggers hanging around the boffins are yoinking the ideas and deploying them for shady purposes. Extending state control, expanding the neoliberal agenda, or the fascist one, or simply trying to turn it in to (more) money.

This can be viewed, and is in fact very neatly viewed, through the ideas of index and representation.

In general, the index is in good shape. These photos are automatically shot, and index whatever they're pointed at. Probably the indexing is a bit shoddy, the quality of these pictures is likely to be lousy, so, while the facial recognition program by god gets a fully indexical thingy of your face, it's done in bad light, it's low resolution, and so on.

Hold on to this point, it's going to become important in a moment.

At this point we have something like representation, but it's of a different sort.

Rather than worrying about how the index (which lies by omission) hits the human/social mind, we worry about how it hits the opaque algorithm. The algorithm works nothing like a human mind. Despite the cries of "AI!" and "neural network!" the algorithm in fact resembles a mind in almost no meaningful way. But, like the mind, the photograph interacting with the algorithm -- call this action machine representation -- can produce real world results.

Let us not lose track of the fact that the algorithm is in general but one piece of a human/machine system. There are always people involved, someplace. The point, though, is that the picture hits the algorithm first.

A friend of mine was driving his vehicle in Pennsylvania, when some cops pulled him over. They were pretty sure the car was stolen, their license plate scanner had pulled up a "stolen car" notification. Things got tense for a bit. Then the cops fiddled with their computer, grumbled, waved guns around, fiddled some more, and then told my friend he could go. He said, because he is smart, "Wrong state, right?" and the cops grumbled some more and nodded sourly.

So what happened here is that the index was perfectly adequate. The numbers were sufficiently rendered, the missing material inherent in the photograph was not relevant. The representation sucked, though. Interestingly, in computer science we also use the word representation to describe how a real thing, like a license plate, might be summarized in a chunk of computer data. So in this case, the representations were misaligned. Either the scanner or the database failed to note the state, and my friend suffered from having a Pennsylvania plate with the same numerical component as a Virginia plate on a car that had been stolen.

The index is the photograph. It is not "Washington State AUC4915."

The photograph, which is just a picture of the ass-end of a car, hits the algorithm. The algorithm them attempts the first steps of representation, and might come up with "AUC4915" or "Washington State AUL4915" or any number of things that are not in fact my license plate number. The action of representation proceeds through the system, representations being processed, altered, and matched against a database that might contain "Virginia AUC4915" or just "AUC4915" and continues with "STOLEN" and then the human part of the system wakes up and starts waving guns around. An arrest may ensue, or not.

In general we're going to see more of this. Face recognition will yield mistaken identity, and the wrong people will be detained, sometimes arrested, and occasionally shot.

You can draw a somewhat shaky line from this notion of machine representation to the traditional one, because in these degenerate times people are doing all this stuff with neural networks, which are trained with sets of existing pre-analyzed photographs. Those photos, the so-called training sets, are analyzed by humans with all the problems of representation that go in to that. Famously, google was identifying black people as gorillas for a while, a case where the generally accepted theory is that representation informed machine representation leading to results. More generally, machine representations are designed by people.

First of all, note how neatly the old and tired theory from the late 20th century seems to be working here, once we disassociate representation from its contemporary day job of supporting identity politics.

So, really, we have three things in play here that seem relevant:
  • Representation, in the human-mind sense.
  • Machine representation, the analog of of representation with the mind replaced with the algorithm.
  • Data representation, the actual structure of database records referred to by the algorithm.

These are all in play, given that the first one tends to inform the third one and, to an extent, the second one.

So what? Who cares and why should we care?

Well, as usual, I think there's value in having things parsed out to teeny little details, because I am me. Secondly, though, this gives us a pretty firm framework for understanding the social/digital mechanisms here. A dash-mounted license plate scanner in a police cruiser isn't a singular object. There are the policemen in the car, there is the database off someplace with the list of all the stolen cars, there are the contractors who wrote all the software, and so on. There is a complex of people and machines involved which, from time to time, produces a traffic stop, an arrest.

The photograph is the input that brings the machine to life, leading to the arrest. In the process, representations collide and interact. If they interact properly, the person arrested in fact committed the crime. If they interact badly, the wrong person is arrested. Or shot.

In computing, we call this systems architecture, but in these cases we need to be roping the human beings and the raw elements of how photographs function into our architecture. We need to understand how these things work. The ideas of the photograph, the index, and the three representations above all need to be grasped at some level. The interactions between them need to be explicitly mapped out and understood if we are to understand where errors are going to creep in, and how to fix them.

And that's just to get the system to work as the designers intend.

We also have questions of whether such systems are good ideas at all. Again, a thorough understanding of what's actually going on, how it works, how it fails and how it succeeds, are useful tools in the arguments against (or for) such systems.

In terms, for instance, of license plate scanning systems we can probably argue that the ship has sailed. The license plate is explicitly designed and intended as a key into a database (a filing system), so arguments that it should not be used as such are probably going to go nowhere. Arguing, though, for correct design, as well as enabling correct design, is still in play.

Perhaps the system my friend was caught up in was designed, in part, by people in a country where there is no state designation on license plates. Perhaps in India, or The Philippines, license plates are issued at a federal level and the numerical content is indeed a unique nationwide key. Perhaps the designed a database record format that did not include the state at all, and instead simply had the plate number indicated as a UNIQUE KEY. When deployed, the American users simply shrugged it off and worked around it, but not well enough.

This would absolutely be a failure of representation, possibly in all three senses. An understanding that representation as discussed in this essay is as much social as technical could have uncover that (hypothetical) problem earlier. An understanding of index might lead to the realization that the photographs weren't good enough to reliably capture the state information, which in the USA can be a bit of a bugger. And so on.

In a way, google's problem with identifying black people as gorillas is falsely comforting. It suggests that if only we have a more diverse community of engineers training our neural networks, then the system will work perfectly.

This is untrue. Machine representation proceeds through a dizzying set of transformations and computations. Google's problem was easy to spot. The sheer complexity, and the sheer mechanicalness, of machine representation is likely to produce far more subtle problems in representation, problems that are deeply inhuman, problems we will not recognize.

Just sitting here mulling this framework around, I can imagine pretty much endless scenarios in which face recognition systems could go wrong, and I can point to exact causes of the failures.