If the AI can't see you, do you exist?

 kw: book reviews, nonfiction, artificial intelligence, artificial vision, face recognition systems, activism, memoirs

One could say it all started with a Halloween mask. One could also say it started with a vision system that couldn't see. Graduate student Joy Buolamwini was coding a project she calls Aspire Mirror, using some off-the-shelf face recognition software, but it couldn't detect her face. She had recently bought a white costume mask for a Halloween event; on a whim she put it on. Lo and behold! The system detected the mask as a face, right away.

Dr. Buolamwini's parents are from Ghana, so she isn't just "Black", she's darker skinned than many Afro-Americans. The software she used for the Aspire Mirror prototype isn't the only one that is blind to her face. Most such software is "Black blind", and many vision systems make many more recognition errors with Black faces than with White ones. Later in her career, when she and colleagues tested major vision systems, they showed an almost perfect correlation between accuracy and paleness of skin. 

As she describes in her book Unmasking AI: My Mission to Protect What is Human in a World of Machines, one major company's software achieved 100% accuracy of both recognition (distinguishing persons who look similar) and verification (admitting only the correct person to their phone, for example) for White male faces only. Other major brands achieved accuracy exceeding 85% or 90%, again for White faces only, and also, better accuracy for men than women. For all of them, the darker the skin, the lower the accuracy of both kinds, and the lowest accuracy is for Black women.

As Big Data has morphed into Large Language Models and Large Image Models (LLM and LIM, respectively), their identification as AI has been cemented in the popular (and press) imagination. Personally, I dislike the term "Artificial Intelligence" when used for such systems. I prefer SI, "Simulated Intelligence". Weaselly terms such as ANI (Artificial Narrow Intelligence) still don't go far enough to distance SI software from natural (i.e., evolved) intelligence. Genuine AGI (Artificial General Intelligence), if it is possible, would be deserving of the raw moniker without the qualifier: AI.

Face recognition is something all animals with vision can accomplish. Of course, various species do better at recognizing their own species. Even insects, with a very different vision system than our own, can recognize one human from another, favoring a familiar face (if it belongs to someone who didn't try to kill them on sight). I saw this when we had a large (basketball size) nest of hornets on a corner of our house. Even in late summer, when hornets get more aggressive, they ignored my wife and me, but would threaten unfamiliar persons; the mail carriers, who change from day to day, were cautioned to avoid that corner of the house. We can conclude that it doesn't take a huge brain to recognize faces reliably.

Machine vision systems have to be trained. So far, it takes a huge "neural network", whether hardware or software based, that has been trained with at least thousands of images of people's faces, to do their job. With millions they do better. The bias problem is with the training data. Dr. Buolamwini found that the standard facial image databases all had more male than female faces, and many more White faces than all others combined. I don't recall much in the way of numbers, but I get the impression that the proportions were "whiter" than the demographics of the American populace. Consider that the major cities of America all have either majority Black populations, or are near-majority Black. That means that automated surveillance systems—for which which most large machine vision systems are obtained—have been trained on a population that they seldom encounter, and have seen too few of their actual "clientele" to be able to reliably recognize them.

The ramifications of this are grave indeed. Scenario: You arrive home to find police or FBI waiting to arrest you. You have been picked out by an AI system as "strongly resembling" a wanted suspect, and a system that was monitoring the several cameras you passed by on your way alerted the authorities of your presence. This has happened to a number of people, not all of them Black, and a few spent hours or days in custody before their actual identity was verified. 

Another scenario: Your company wants to reduce the number of security guards at the gate (I've worked at such secure installations), so they announce that you'll need to look into a camera to get access. If the system being used is no more capable than the one behind Aspire Mirror, and you happen to be Black, it may not even recognize that a face has been presented to it!

Another issue that the author presents, several times: Training images are typically frontal and partial profile photos with good lighting and a neutral background. Nothing like pictures taken "in the wild", such as from a camera at the top of a 20-foot pole on a street corner.

This reminds me of something I'll have to describe, because I can't ethically show the photo here. After a baptism in my back yard, we took a group photo of about 35 of us. We were a good mix of races including two Black families and a few Chinese families; five Caucasians were present (I am in a very multi-ethnic church). It is under a tree, so the lighting is dappled. I did brightness tests of several faces. One White boy's face in a shadowy part of the image is darker than the face of a Black boy in the sunshine. A Black man in the shadow is practically invisible. However, shifting the photo's brightness a little makes that particular Black man easy to see, while washing out all the sunlit faces. Interestingly, Google Photos seems to be pretty good at isolating faces and recognizing them; it recognized every face in the photo, even though it had to ask me to verify some of the identifications. Other face recognition software that I've used don't do quite so well.

This example emphasizes that even a posed group photo can have lighting variations that stretch a vision system (including our own!) to its limits. With such matters as the prime example, the author expands the arena to include automated systems of decision support: Résumés scanned and prioritized before the hiring manager even sees them; street surveillance with automated flagging of "suspicious" persons or behavior; neighborhood analyses that set store prices based on demographics, and similarly for real estate appraisals. All these are actually extensions of things that humans were doing already. The machines just do them faster and more cheaply...and some have been banned as a human activity (real estate redlining, for example), but AI systems are being allowed to slip under the radar.

Only a portion of the book is devoted to such technical matters. Much more involves Dr. Buolamwini's increasing involvement in policy. As she relates, it is not enough to point out the deficiencies of systems that lie behind automated decision making. Their weaknesses reflect and even amplify the weaknesses of the people who create them. Biased people cannot produce unbiased systems (A side thought: Get together several people who each recognize their own bias, and ensure that multiple viewpoints are included, and they have a better chance of reducing the biases of a system they all have a hand in developing and training; this is hard! But it is analogous to the way TCP/IP protocols can transmit data with extraordinary perfection through a noisy network; the Internet would be impossible without it).

Over a rather short span of time, much of it while she was still in graduate school, the author has become a more and more visible presence to policy makers, in her capacity as a founding member of the Algorithmic Justice League and the Poet of Code. In her epilogue she describes a meeting with the President of the U.S., at which he asked, "What should I do?" 

Since this occurred quite recently, it will be left to the next President (and maybe several "nexts") to deal with these matters. Make no mistake: they must be dealt with. This is a bipartisan matter. "Algorithmic harms" know no political party, and oxen all along the political spectrum have been gored. I find it abundantly clear that long before AGI arrives to become just "AI", making all prior systems obsolete, governance systems are required to rein in the big money behind overuse of such tools. We still have a grip on the tail of this tiger, but it is quite capable of spinning 'round for a bite!

Cities where no city should be

kw: book reviews, nonfiction, activism, city planning

About seven years ago, in the aftermath of Hurricane Katrina, I wrote to the governor of Louisiana, suggesting that the funds being earmarked for rebuilding New Orleans would be better spent relocating the population to higher ground, possibly north of Lake Pontchartrain. I never heard back, directly, but I did hear her in a press conference two days later, very angrily dismissing suggestions that the current site of New Orleans ought to be abandoned. Thus the power of tradition over common sense.

I have similar concerns for Phoenix, Arizona, which has nearly five times the population of New Orleans, and the greater Phoenix area has about 12 times the population, or more than four million people. I have visited Phoenix a few times (rather unwillingly after the first visit). I don't understand why it is there.

I am not the only one. The title of the book yields a clue: Bird on Fire: Lessons From the World's Least Sustainable City by Andrew Ross. Folks may quibble whether Phoenix is really the least sustainable. But the idea of locating 4+ millions in an area in which nature provides for no more than 40,000 is insane. However, that is not the central thesis of the book.

Prior to 1860, between 10,000 and 20,000 Pima (or Akimel O'odham) people lived a genuinely sustainable, agricultural life along the Gila and Salt rivers. Their life was probably similar to that of the Hohokam culture that once flourished in the area, and may be a continuation of it. Then in 1861 Anglos set up a city where Phoenix is today, atop the ruins of abandoned Hohokam canals. A generation or two later, once the Anglo population began to grow rapidly, they dammed up and siphoned off the waters upstream of the Pima, and essentially drove them into dependency by artificial drought.

Now Phoenix is one of the nation's largest cities. What has been behind the huge increase in population? First, huge chunks of the "military-industrial complex" about which Eisenhower warned us, taking advantage of cheap land and a measure of insulation from the more populated (and thus more watched) areas in the Northeast. Secondly, rampant and short-sighted boosterism coupled with an extreme frontier "don't tell me what to do with my land" mentality. And, most telling, Colorado and California and Nevada let them get away with appropriating out-of-state water supplies such as about one-third of the Colorado River's flow, as "defined" during an exceptionally wet period.

In Arizona, more than any other state, regional politics is dominated by real estate speculation. A significant dent has been made in this picture during the past five years, but the mentality remains (See yesterday's post: former half-million-dollar homes are now hard to sell for $200,000). Plans are still being discussed to develop 275 square miles in the foothills of the Supersition Mountains; there is some chance a more reasonable plan than the "sprawl" that surrounds the area will prevail, primarily because there is a recession going on! Arizona has the highest proportion of people with "underwater" mortgages in the nation.

The author interviewed hundreds of people for this book, and the nearly uniform symptom was that any troubles were somebody else's fault. Those in charge of water policy cannot imagine why anyone should object to central Arizona's demanding increasing allocations of water from elsewhere. Those supposedly watching over the environment just throw up their hands about the fact that most of the groundwater under the county is too polluted to use (courtesy of those military-industrial plants, now mostly closed down). Hardly anyone in authority is willing to listen to any plan that doesn't include nearly doubling the population in the next 50-80 years. They just bemoan the "white flight" elsewhere that has led to a rapidly increasing proportion of Latinos, even though many Latinos are leaving also.

Is it really true that capitalism requires a growing economy to work? I certainly hope not, because sooner or later every economy is certain to contract, no matter what the population is. There is only so much water, only so much coal or oil, only so much land to farm or to live on. The Phoenix area is a look into a future we all face as the various amounts of resources we depend on become more and more scarce. One subchapter is titled, "Building homes for the home builders." The "plan" of eternal growth is actually a Ponzi scheme. It is just most evident in Arizona.

I have a friend in New Jersey who tried to make a living as a farmer with a 25-acre farm. He has since said he really needed 100 acres to make a proper go of it. The yearly rainfall in the area is about 40 inches (one meter). Prior to moving East, we lived nearly a decade in Oklahoma. The smallest viable farms and ranches there are in the 500-1,000 acre range. Yearly rainfall is a bit over 25 inches (0.65 m). Prior to that, we lived in South Dakota. My next-door neighbor was one of the few successful ranchers. Another friend had a mixed farm and ranch a couple hours' drive away. Both had spreads larger than 10,000 acres. I remember getting a detailed dissertation from my neighbor about why it required 50+ acres of prairie grass to support each cow, and how he had to grow a certain amount of alfalfa for hay to be stored against the occasional drought. A 10-15 year accumulation of hay could be devoured in one bad season. The rainfall on Table Mesa in South Dakota is 12 inches yearly (0.3 m).

Central Arizona gets 8 inches (0.2 m) of rain yearly, as averaged since the 1980s. Recent years have seen less rain, and climate warming is likely to keep the rainfall in the six-inch (0.15 m) range in coming years. Even in a good year, Arizona is below the threshold of widespread agriculture. A cow can't walk far enough to find the food it needs, and will starve, except along the few rivers. The relatively small numbers of the Hohokam and later the Pima were all the land could support, and then only within a few miles of those rivers.

In the Mojave Desert, as a Geology student I learned that a swimming pool will evaporate five feet (1.5 m) of water yearly. In central Arizona and the rest of the Sonora Desert to the south, evaporation exceeds 6.5 feet (2 m). Damming up a river to make a reservoir just reduces the total usable water from that river.

Mr. Ross hopes some of the efforts now under way in Arizona will show the way for improving the sustainability of other cities. One hopeful sign is the recent culmination of an 80-year legal battle that gave restored water rights to the Pima people. Some of them still know how to make good use of that water for something besides a Kentucky bluegrass lawn or a swimming pool (There are a lot of swimming pools in northern Phoenix). But I say there's not much wrong with Phoenix that an out-migration of several million people would not cure.

Half a Superbunny autobio is better than none

kw: book reviews, nonfiction, autobiographies, activism, animal rights

For years, I've wondered, "Just what is PETA anyhow?" I happened to see the back of this book first, and once it caught my eye, it didn't take long to find that the Rabbit-Man is none other than Dan Mathews, a VP at PETA. The book is Committed: A Rabble-Rouser's Memoir.

Dan is a master of getting attention. Of course, since just about day one in school, it was usually attention of the unpleasant kind. His schoolmates had him pegged for a "fag" long before he adopted the label. Now he's almost as active in gay rights as he is for animal rights. But he took the lemon and made lemonade. He's had a couple decades to hone his craft, which now consists of three steps:

1. Attempt a civil approach with a target person or organization, whether Calvin Klein or KFC. This is usually fruitless.
2. Subject a number of said target's employees, including the CEO or other honcho when possible, to an amusing, outrageous slice of Hell for a short period (minutes, hours, sometimes days). This sometimes prompts a change of direction...or at least of things said about direction.
3. In rare cases, someone like Calvin Klein will send a message, "Why didn't you just ask to see me?" and the answer is, "We really, really tried with [documented] results." The ensuing sit-down usually gets the best results.

Dan's earlier "work" was more heavily weighted on step 2, until he learned "that apathy and indifference are more easily conquered with charm than antagonism", as he writes on page 56.

However, institutions and important people are often very, very insulated. It can take a lot to get their attention. Getting attention is what Dan Mathews does best. Whether posing as a big rabbit, or carrot, or as a priest, in formal dress or none whatever, he gets attention.

I've been quite put off for years by the PETA tactics. Now I understand a little better. Almost by happenstance, PETA performed two "actions" on behalf of a Harvard class. First, they did a version of "I'd rather wear nothing than furs". That got Dan and one or two others arrested for indecent exposure. Then, taking one student's suggestion, they had a much more civil demonstration. The first event was seen by millions; the second, by scarcely a dozen. The students got the point.

It is a pity such tactics are required, just to get someone to sit still to hear a complaint. In the face of some real, and really dangerous, fanatics out there, PETA is actually rather innocuous. They are also useful. We need our gadflies, our finger-pointers, our loud 'voice of conscience' activists, to give civilization at least a ghost of a chance at being truly civil.