Tuesday, March 28, 2017

Facebook, Fake News, And The Problems Of Trust And Legitimacy

When I heard that Facebook was going to try to "do something" about fake news, my first thought was "Oh yeah, that'll end well. What could go wrong"?

It's not that lies, hoaxes, and misinformation aren't a problem (and by the way, what was wrong with "lies" "hoaxes" and "misinformation"?) -- it's just that truth and factuality are not simple problems, they're not algorithmic problems, and they're not problems you can take a "neutral" stand on. Yet you know Facebook will try to treat them as if they are.

The first kind of example that came into my mind on thinking about this was about the aftermath of the protests in Ferguson after the shooting of Michael Brown. I got a lot of my news from following people on Twitter who were there -- some, like Jelani Cobb, professional journalists, and others people who were not. 

Often, what I read on Twitter did not match up with what I read in the mainstream press. The press interviewed the police or asked government officials to comment on things; in the nature of things they had a vested interested in portraying the protestors as initiators of violence. Reports from people there emphasized the militarized police response and also the number of peaceful protestors doing things like protecting property and cleaning up.

When the reports of the citizens on the ground don't match up with official reports or reports in the news, who are you going to trust to tell you the truth? And when, and why? These are difficult questions. But do you really want Facebook answering them for you?

Maybe you might say the is for simpler, more straightforward cases (like, you know, "lies," "hoaxes" and "misinformation"). But that's not how the response to "fake news" has been shaking out so far. Maybe you've heard about the "B. S. Detector" that claims to "alert users to unreliable news sources." One of the first things that happened was the site Naked Capitalism got incorrectly tagged as a "fake news" site. In fact, what Naked Capitalism is is in-depth analysis of current events that sometimes diverges from official positions and mainstream media.

Are we really so far down the rabbit hole that we want social media companies to pronounce on what is and is not a legitimate critique of government statements or the New York Times?

Facebook, in fashion characteristic of the tech industry, wants to be address the problem of "fake news" while also maintaining "neutrality." As we've discussed before, the dream is to off-load judgments onto users so that algorithm's can solve all problems and no value judgments have to be explicitly endorsed. And as we've discussed before, this is impossible: there is no "value-free" way to offload judgments about what is and is not acceptable speech, or what does and does not constitute unacceptable forms of discrimination, or what is or is not sexist, racist, and so on. You let users decide you're often going to get an outcome that goes horribly wrong.

I had to laugh when I learned that the term Facebook is going to use for hoaxes, lies, and misinformation is "disputed." For one thing, could anything be a more obvious attempt to sound "neutral?" It's like, "WE'RE not saying there's a problem. But SOMEBODY out there is disputing this."

In a more sinister vein, when it comes to actual hoaxes, lies, and misinformation, doesn't "disputed" actually seem like it would add an air of legitimacy? One of the more interesting things I read about "fake news" was how that guy in California created a ton of fake news  -- like "FBI Agent Suspected In Hillary Email Leaks Found Dead In Apparent Murder-Suicide" -- and earned a ton of money. You can read an interview with him here. But then it turns out that teens in Macedonia (and presumably people all over the world) are creating fake news just for profit.

Isn't using the term "disputed" to describe "FBI Agent Suspected In Hillary Email Leaks Found Dead In Apparent Murder-Suicide" heading off in the wrong direction entirely? Doesn't that make it sound like some obscure point about Benghazi, where some partisans think one thing happened and some partisans think another thing happened, but who really knows? Doesn't it make it sound like a possibly legit thing? When, in fact, it is just a hoax, a set of lies!

The problem of truth and belief goes way beyond algorithms and neutrality and involves complicated issues of community and trust. When the New York Times ran a whole article explaining why "pizzagate" was based on a set of lies, do you think people who believed in pizzagate said to themselves "Oh, I guess that pizzagate wasn't true." Of course not. They went and wrote articles debunking the debunking. Just a few days ago there was a protest in DC with people demanding an inquiry. 

Looking up pizzagate on Wikipedia, I see a journalist quoted as saying that pizzagate is "two worlds clashing. People don't trust the mainstream media anymore, but it's true that people shouldn't take the alternative media as truth, either." This is aptly said. People trust different sources. No algorithm is going to deal with that problem.

If Facebook's proposed solution is to add a "disputed" tag to posts, potentially undermining citizen reports that contradict official news, and legitimating things that are lies and hoaxes in the first place -- well, it seems to me this may well do more harm than good. Maybe Facebook should stay out of the social epistemology business altogether.

Tuesday, March 21, 2017

The Modern Condition: No Slack In The System

There was maybe going to be a transit strike this past Monday in the area where I work, starting this Monday. It didn't happen, because of a tentative agreement reached late Sunday night. But it got me thinking about the ways that systems of modern life are so tightly wound that any disruption is like the end of the world.

Of course, having a major bus system stop functioning is a big deal no matter what the time and place. But it feels like these days, especially, a lot of people have employment that is particularly inflexible, precarious, and high-pressure. Work schedules can be posted late. Failure to obey the schedules can result in reprimands and dismissals. Quotas are set for a range of criteria and if you can't meet them, well -- you're screwed.

I don't know if you've been following the news about the Amazon warehouses, and how the pickers are under constant surveillance, not allowed to sit ever, forced to aim for a "target rate" of 100-120 items an hour. This story describes how Mail UK deals with employees as independent contractors, so that if they get sick, they not only don't get paid, they have to pay for replacement workers; a worker was charged £216 per day of absence after got hit by a car while delivering packages. Bankers across Canada are told if they can't upsell enough products to people who don't need them, they'll be fired.

But it's not only labor where there's no slack in the system. If you ever fly these days, you know that if something goes wrong with your plane, or a crew member gets sick, or there's bad weather or whatever, there's no "Oh, we'll get another plane' or "Oh, we'll put you on the next one." The planes are all in use; the crew are all maxed out; the planes are all full. There is no duplication, or overlap, or plan B, or whatever.

One thing about this that interests me is that although I have used a negative formulation to describe the phenomenon I am talking about, there is another description of the same thing, a positive one, one you'd probably find more often in the Business Section of the paper, and that is: "It's efficient."

It's efficient in one ordinary sense of the word: you're doing as much as you can with the "resources" you have. Amazon moves a ton of stuff for low financial cost. Planes fly a ton of people with lower fares. UK Mail made a profit of £16m last year when it was bought out by the Deutsche Post DHL Group.

You can ask the question of why "no slack in the system" seems to be so dramatically on the rise, but once you notice that "no slack" is also "financially efficient," you start to wonder about other things, like why this didn't happen earlier, or why there used to be so much flexibility, easy-goingness, and duplication, or why this is all happening now.

To these questions I do not have answers. Is it that electronic communication made possible a tightness that wasn't possible before? Is it that globalization and the financial crisis made everyone focus their attention on the bottom line? Is it a cultural thing involving negative attitudes toward labor and consumer protections? Or maybe it's actually been a really gradual thing that just seemed dramatic to me?

There's a point of view from which an important part of the explanation of things like this involves "corporate greed." The idea is that in a normal world, corporations are happy to make a moderate amount of money, and prioritize other things like worker well-being and so on. So the problem is that "greedy" corporations are trying to make a lot of money, instead of a moderate amount of money. And so they can't prioritize anything else.

As I've explained before, I think this explanation is inaccurate and possibly naive. In a modern capitalist marketplace context, the pressures toward efficiency are enormous. If you're less efficient, you'll just get run out of town by some other organization that can offer the same product for a lower price. In fact, this is just what we've seen over and over again, with smaller retailers going out of business because Amazon, Walmart, and so on are so hyper efficient. So: often it's efficiency or die.

I don't know whether we ought to do anything about the slack-freeness involved in things like having no planes sitting around unused. Fewer and more packed airplanes is actually better from the environmental point of view.

But when it comes to workers, my sense is that the slack-free workplace is horrible for people. It creates jobs that are massively stressful and ruin people's health and well-being. It illustrates something we've talked about before: that what is efficient when you're measuring money is not always what is "efficient" at producing good outcomes overall -- assuming "good outcomes" includes personal well-being and happiness of people.

Given the competitive nature of capitalism, it seems to me any solution will have to be systemic, and will have to involve labor laws, worker organizations and so on. Given that the bus driver's union Unifor Local 4303 retweeted a link to this webpage, about fairness in labor laws, including a comment about how "fair work schedules" means "2 week's notice," I'm guessing they're thinking the same sort of thing.

Tuesday, March 14, 2017

Humanities Teaching Is Difficult And Time-Consuming

Before I studied philosophy, I studied math. I was working on a PhD in set theory when I became less interested in how-do-you-prove-the-theorem and more interested in questions like what-does-it-mean-if-you-can't-prove-or-disprove-the-theorem?

Among other things, this means that before I started studying philosophy I spent a certain amount of time teaching mathematics. I started as a teaching assistant for introductory courses like Calculus and Statistics, and then I was a teaching assistant for more advanced courses like Differential Equations, and then toward the end I taught a few classes myself, including one on how to do mathematical proofs.

Math is hard. But I found teaching mathematics to be mostly straightforward and rewarding. Students are usually externally motivated to learn: they want to do physics, or engineering, or more advanced math, or whatever, and to do it they have to learn some math. Except in the case of bogus requirements -- like baby calculus for no reason for majors like business to "weed out" students they didn't like  -- the importance and relevance of the subject was relatively obvious.

At the level of undergraduate teaching at least, math is also coherent and unchanging. Because of the nature of the subject, the same kinds of things confuse people, and similar kinds of questions arise again and again. Once I had explained concepts like limits, differentiation, and integration a few times, the ideas were cemented in my head in such a way that very little teaching preparation was required.

On top of everything else, because math is obviously difficult, a teacher's ability to break down difficult concepts to make them seem simple earns them great respect. And this was something I was relatively good at.

Several years into the process of studying for a PhD in math I switched to philosophy. I've now been teaching philosophy in one form or another for ... well, a lot of years. And my personal opinion is that teaching philosophy is way more difficult and way more time-consuming than teaching mathematics. I don't have a lot of experience with the other humanities, but it is my belief that the reasons apply to humanities teaching generally.


Those reasons are several. For one thing, math seems difficult and a teacher is there to make it seem simpler, but in the humanities, it's often necessary to start by taking something that seems simple and showing students how difficult it is. I teach ethics, and philosophy of sex and love, and contemporary moral problems, and philosophy of economics. In all of these areas there's a sense in which a student already knows what they think about things, and part of my job is to complicate that -- to raise questions about things that seem obvious, to showcase views that seem counter-intuitive, and to just generally show how many different factors and perspectives can come into play.

This is intellectually difficult, and it can also be emotionally draining. How do you frame the issues when students are coming into the room with very different background assumptions - and you don't even know what those background assumptions are? How do you encourage people to speak up when part of your job is to suggest they might be totally wrong? How, exactly, do you figure out the line between constructively challenging existing beliefs and just being a contrarian pain in the ass?

Some people love the way humanities thinking challenges them, but other people find it exhausting and annoying. Sometimes science students in my ethical thinking class tell me how frustrated they are by the lack of a "right answer" in philosophy. I sympathize! It can be frustrating as hell. Unfortunately, the problems we're talking about are the ones that don't have straightforward answers, so it's the best we can do to muddle through.

Another factor, of course, is the variation and unpredictability of what kinds of things are going to come up. The social and cultural world we're living in makes different things seem obvious in different times. Even just contingently there are classrooms where one thing seems really important that didn't seem important to some other group.

This variation and unpredictability is, of course, part of what makes humanities teaching so important, relevant, engaging, and fun. But it also means that while teaching an interactive mathematics class can feel like a going through a play you're performed a thousand times, teaching an interactive humanities class can feel like a high-wire act where the tricks are constantly changing.

And finally, of course, there's grading. While mathematics grading can be time-consuming (when I did it, we didn't just grade yes-or-no, we looked at student work for partial credit) it's not like grading a paper -- work that combines engaging with someone's novel ideas and helping them toward an amorphous goal like "writing well." As we've discussed before, it takes a lot of time and energy, and it's not something you can scale up.

A few times recently I happened to be in large university group settings, where people were coming from a range of disciplines. And in that context, I heard some remarks about how, from the point of view of the sciences, what we humanities might regard as a large-ish class -- like, 50 or 100 students -- is to them a very small class. No one said it, but I felt the suggestion that somehow we humanities people weren't pulling our weight, that what we were doing was some kind of niche thing, cute and nice if you can afford it, but not really where the action is.

And I can't really say, because I have no experience teaching science. I only taught math -- which to me is a completely different kettle of fish. But from my perspective, the time and energy to teach a philosophy class is way more than the time and energy of teaching a mathematics class. Even when the classes are a lot smaller.

None of this is meant as a complaint. I love university students, and I love being around them. I think the people who criticize the younger generation for being phone-obsessed and jobless are wrong and ill-informed, and that today's young people are the hope of the future. I regard helping these young people understand the complex world around them as one of the best things anyone can do.

I'm just saying: for me, anyway, teaching about utilitarianism is way harder than teaching what it means to take the limit as h goes to zero.

Tuesday, March 7, 2017

Nudge Versus Trust In Modern Politics

Maybe you're encountered the concept of "Nudge," in which framing choices in various ways can push people toward doing certain things without directly forcing them to do those things. In a relatively innocuous example, you might notice that in a cafeteria setting, people choose more vegetables when vegetables, instead of desert, are placed right up front, and as the cafeteria director you might decide that is a Good Thing.

Nudge grows out of the theory of behavioral economics. Behavioral economists noticed that not only are people irrational, they are systematically irrational -- they fail to do things that seem in their otherwise best interest, because they don't have great impulse control, or because they're not good reasoners, and so on -- and they fail in predictable ways. Nudging exploits that predictability to shape outcomes.

I consider myself a progressive and a lefty, and you might think that this political orientation and the idea of nudging would go hand in hand. Progressives want to bring about change on big complicated things like protecting the environment, where collective action is really difficult -- maybe nudging people toward energy conservation would be a good idea? Progressives often see people as influenced by context and culture, rather than as atomic and autonomous individuals creating their own self-made way in the world -- if people are affected by context and culture, why not try to make that work for us rather than against us? And examples like the cafeteria speak to me. I am just the kind of person who wants to be nudged toward eating more vegetables, and I'm just the kind of person who recognizes that nudging could work.

But nudging is often creepy. For one thing, it's described as "value-neutral" -- nudgers are just helping you do what you would do if you weren't so systematically irrational. But this is just implausible. How do you know what what people would do if they weren't so systematically irrational? In fact, you have no idea.

As we've discussed before, the person who eats a lot of desert may be irrational. But they might also be rationally satisfying a strongly felt preference for cake over the things that you get from foregoing cake. People have priorities other than living longer, and as Paula Poundstone says: What part of Ring-Dings make my life worth living do you not understand? As has been pointed out (e. g. in this book by my friend Mark White), the risk is that policy makers are projecting their own sense of what matters onto the situation.

As time went on, and the more I saw nudges in action, the more suspicious I became. And then a couple of weeks ago I read this New Yorker article about the use of behavioral economics and nudging in the context of the Flint water crisis. When I first saw the topic, I was like, WTF? The people who behaved badly in the Flint water crisis weren't the citizens. They were the government agencies and representatives who made a terrible decision to divert the water sources to save money, then covered it up and lied about it, then blamed one another, then failed to do anything to fix it. Was the author going to talk about nudging top-level decision-makers? Now that would be interesting!


No, of course that wasn't it. The article was about how you could create structures that would get people to do things like get and believe up to date information, and act on that information by doing things like changing filters and so on.

I guess filter changing reminders are reasonably innocuous in the circumstances. But, as the author of the article kept bringing up, the real problem between the citizens and the government Flint wasn't about information and facts and "rational behavior." It was about trust. The citizens didn't trust official representatives to tell them the truth about the water situation. And FFS, why would they? They'd been lied to and manipulated from this end to that. And now someone shows up saying they're from the government and they're there to get you to do some things rather than some other things and believe this thing rather than that?

Reading this story I just felt what a profound disconnect there seemed to be between the nature of the problem and the proposal on offer. Trust was destroyed. And the situation in Flint is still fucked up. And yet people want to use brain science to figure out a "strategy" for getting the citizens to do one thing rather than another? 

There's a fascinating exchange toward the end of the piece, when the behavioral team meets with a local activist. There are immediate cultural disconnects -- like where the activist offers to share some special fried chicken and the team members decline because they're vegetarians. But eventually the activist asks the team to just talk about how they're feeling since the election, and the main point person sort of breaks down and talks about how shitty and frightened she feels. And this is what creates some connection, and some trust, between them. Because it's people being honest with each other.

I think trust -- and its erosion -- has played an important and complicated role in a lot of recent politics. Obviously people have started getting their information from different sources, and it's often been noted that the body of shared facts and background we can all rely on in talking with one another is getting smaller and smaller.

This is often described as if some people, and not others, are just failing to judge in accordance with the evidence, willfully ignoring the facts in favor of their own opinions. And sure, there is some of this going on. But at a deeper level the problem is also a problem of trust. People don't trust the same sources as sources of evidence and facts, and they don't trust each other. 

It's frustrating to me to hear people talk about the lack of shared belief like it's a relatively straightforward (if difficult) problem -- where people just have to be brought around to proper belief formation. Maybe we can all get better at how we figure out what to believe. But this start to this process won't be a top-down nudge style step. In fact, a top-down nudge-style step can be just the thing that erodes trust, evoking, as it does, an I-know-the-story-and-you-don't kind of mood.

I don't know what the answer is, but I think it will have to involve openness, honesty, and maybe even mutual vulnerability -- things that nudges don't really have anything to do with.