CHAOSS CHAOSScon EU 2020, 22 Feb 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keynote: Ethics: What You Know & What You Don't Know | Deb Nicholson | CHAOSScon EU 2020

Description

Keynote: Ethics: What You Know & What You Don't Know

Metrics give us a lot of insight into what's going on with our projects -- but only if we have enough context. The catch is that personal biases and pre-conceptions are extremely hard to filter for so no one person can see everything. Those gaps means that even the most well-meaning of us could end up doing more harm than good. In this talk, I'll share a few examples of the ways human brains make assumptions, some of our typical blind spots and how incomplete data can cause harm when deployed without any checks. For folks who collect, analyze and ultimately make mission critical decisions with metrics, the challenge is to constantly maintain a bit of healthy skepticism of what the numbers seem to be telling us.

Slides: https://chaoss.github.io/website/CHAOSScon/2020EU/slides/ethics.pdf

-------
About CHAOSScon

Learn about open source project health metrics and tools used by open source projects, communities, and engineering teams to track and analyze their community work. This conference will provide a venue for discussing open source project health, CHAOSS updates, use cases, and hands-on workshops for developers, community managers, project managers, and anyone interested in measuring open source project health. We will also share insights from the CHAOSS working groups on Diversity and Inclusion, Evolution, Risk, Value, and Common Metrics.

https://chaoss.community/chaosscon-2020-eu/

A

Your problems and urethanes and what you're trying to get.

A

A

And then we make email.

A

And you know so we're like oh we're, gonna like make better brains, but like we barely understand the brains that we have so making one. That is a machine that does like what we think our brains do is interesting, because we don't even know entirely always understand what our brains are really so simply there's an entire book about this. The human brain is a is kind of a clutch.

A

It's a mixed match like how your your right eye reports to the left side of your brain and vice versa, playing set has something to do with birds. It's a long story. The whole body is like that, though, like our spines, you could not design a better spine for back problems. So all of it is like that. So so the way we're wired to think about resources like sometimes we get really focused on numbers and then sometimes it's like.

A

Have you ever driven like 2 or 3 hours to save like 20 or 30 bucks and you're like ok, you don't have to admit it or have you ever been really mad at the person who booked you a 5 a.m. flight and it was you.

B

C

A

Like hundred dollars cheaper, so we do those kinds of things all the time we get like focused on this number and then we're like oh yeah and then you're like I got a P have been in the car for two and half hours. Why you might save in 20 bucks, it's ridiculous. So so that's one thing: that's really funny about the way that our brains, work.

A

Another man is interesting is the way that we make memories like, interestingly enough, the persistence of memory, but almost everyone forgets the name of it and calls it like milky clusters and Sergeant.

B

Russ, alright completely forgot the mic. I'm.

A

Loud anyway, so but.

B

You're recording just trying to make it available.

A

B

A

That alright, um so memory is actually really strange. There are all these stories about people having a distinct recollection of having done something that if you go back and look it's impossible, like a really famous example, was we look like there was folks who ask people about their experience on 9/11 the day that the twin towers fell in New, York and people had these distinct recollections and then it was like wait. You didn't you're living in Queens them, there's no way that you were like across the street.

A

This is ridiculous, but because the bring like layers, things in, like you, remember things that actually didn't happen, and so so I guess what I'm saying is we're not perfect. So when we try to do things with computers or like with metrics and get a clear real picture of the truth, we're starting with some imperfections already built in and the key is to not like perpetuate the mistakes right. So this is, if you can't see it, it says: Facebook, grammar friends, did you mean Facebook, G, ramm, er, and so enough people have misspelled grammar.

A

This way that now, when you go to Google, like you know and I, don't know if this is still the case, maybe they fix it, but it will suggest the bad spelling that many many people use so so this is. This is sneaky like it's.

A

um You know like we're, trying to train computers to be like us, which might not be the best go hobbit side, um and so you know it's but like the you know, dolphins are not stepping up to the task, so I guess it remains to us to program computers and have them do things um so I want to talk a little bit about stereotypes and one so I book whistling Vivaldi and I was a number of generally interesting studies, but one thing is that we're actually capable of stereotyping ourselves, which is super weird right.

A

So um so they didn't test, and this is in the US and they asked asian-american women to take a math test and on the first test they were grouped with other women and in that second test they were grouped with other Asian Americans. So the stereotype is that women are bad at math all right and so and then there's another stereotype that asian-americans are good at math.

A

So what happened was that asian-american women performed more poorly on the math test, a lot they didn't tank it, but like a little bit more poorly when grouped with other women, because they were like. Oh, the point of this study is to see how women are at math and I. Guess you know we're bad at it so and then the other one they were like.

A

Oh, the point of the study is to show like how asian-americans are good at math, so obviously like I'm good at math, so just believing that you were part of a group that was bad at math or good at math affected the results.

A

So when I say that, like you know, when people talk to you about when you're, taking notes or you're calling a group by a certain name or you have a question that is weird or leading like when they say it matters it matters like you already have it inside the ability to stereotype yourself. So it's very easy to stereotype someone else without being aware of it. I, don't think that many of us are aware of stereotyping ourselves. We just do it so um another thing that's interesting.

A

This one is a fictional example is timing like sometimes it's very easy to like assign significance to the timing of something. Has anyone read the book Dahlgren by Samuel Delaney, all right start out to get this right because no one's read it. It's a great one.

A

So it's said in this like near future a post-apocalyptic world for some reason, it's probably approximately Chicago. For some reason the city's been cut off and they don't have electrical power anymore, and no one knows why there's a million conspiracy theories going around and like why they got cut off what happened: bla, bla bla, so everyone is suspicious and stuff all the time. We start to get things done, so um so the main character, the kid he never gets.

A

Another name has roommates like you do, even in the even the you know, apocalypse in the apocalypse you're still gonna have roommates, and- and so he had this glass of water like so. An interesting thing about modern society is that water that comes from the tap is typically has bubbles in it. The reason and like sometimes you'll, see them, sometimes don't, depending on how strong the bubbles are, and the reason is because anaerobic bacteria can live in the water.

A

So if you put water, if you put air in it, it kills any anaerobic bacteria before you drink your glass of water, so yay, so he started noticing that uh the glass of water who was drinking each day had fewer and fewer bubbles in it, and so he had all these theories he thought like they like was it. The government was an aliens. Was it like.

A

Some rival city was like messing with the water, and it was like a little bit each day like a little bit fewer bubbles each day, and then he saw his roommate doing the dishes and his roommate was doing the whoops poppin job on the dishes. It wasn't the water at all. It was just a roommate doing a more and more haphazard job of washing the glasses.

A

So but it was like, oh you, because he was primed for conspiracy and like looking for like them messing with us and all these kinds of things like you immediately jumped like the timing means like obviously now they're messing with the water. So um you know so like it's. It's uh it's easy to assign timing like outside significance.

A

Well, alright, uh one of these guys got messed up, but okay, so um so history and context X, don't. Why is what that other side says? Oh a lot of these are messed up. um So um sorry, my friend in the US and like a number of years ago we were talking- and she said, African Americans, don't swim and I thought, oh well in America and you're telling me that she don't swim and, um and then I had heard it from also from like comedians.

A

It's like a joke in some ways results are not a joke about 70 percent of African American children swam or can only see them a little bit and then occasionally people drown because they can't swim, and so you might think, like. Oh that's kind of like a weird thing that African Americans know. So maybe it's cultural like how like I'm a white person Protestant and my parents like martinis and meatloaf and hate spicy food, and this is just how they are like.

A

But if you knew the history and the context, then it might be a little different. So if you ran a community pool, you might be tempted to be like well, we don't have to like asking you questions about whether African Americans are showing up because they don't swim.

A

But if you understood the history, the swimming began in the US and like public pools in the 1920s and 30s and they were segregated and our Center is sometimes with police force and and violent interactions. So Sammy, Davis, jr. famously went to a pool in Las Vegas in 1952, and he was famous enough that they didn't mess with him, but they drained the pool after he got out.

A

So that was like you know it was this really weird thing so um so African Americans don't swim, just not because they're like in pools, but because there's this violent history of enforcing segregated pools, so um we're gonna owe justified, okay and and that continues today this person. This is just from like two years ago that uh it's like a apartment complex with a pool there's a black woman in his pool and he caught up.

A

She was in the building, but he'd never seen her because they're well they're, not friends, and so he called the police and and put someone caught him and like talking about how, like you know, he's supposed to know who's in the pool or whatever, um and obviously it's not true that African Americans don't swim. This is moon actually Manuel and she won an Olympic medal for swimming gold in 2016. So um if you find yourself cutting a corner like oh, we don't have to count this group or we don't have to this metric.

A

You might ask yourself: why does everyone cut this corner? You may be contributing to like a historical inequity. They may be helping to continue a history of violence and people being excluded from things that they would want to be at, and so so you know when you think about the categories that you make and who you count and what you count, maybe tranquilo more like mr. Rogers, and it's assumed that inclusion there's a famous moment on that, mr.

A

Rogers house, and so if you don't watch that in Europe and I mean I know the us kids. It was like kind of inescapable, which is great, um but businesses friend officer Clemens, who he decided to share the pool with on TV, and it was a very specific moment so um so yeah I, maybe try and be that spirit. Just assume that your friend wants to join you in the pool, whether it's your community pool. Maybe the pool is a metaphor for your community.

A

So the power of categorization is so when you're deciding to trash different things you get to put them into the categories right. So um a couple of these grabbed a different slide background um anyway, so this person wrote a book called Paris and other disappointments right. That's a really weird category right! It's like you can pretty much assume you did not enjoy Paris it's expensive! There's a lot of stairs I think it outweighs that you know the croissants and stuff outweigh that, but he didn't think so.

A

So if I wrote, if I wrote a blog post aside like Python and other things that make you want to stab yourself in the face, oh you don't like Python, okay, so they you know or or something like that. So when you weak like I, hope that we don't put things in our metrics like start contributors and jerks who left you know like that is very far from neutral. Obviously so um so the way that we choose categories it, it makes a huge difference and a lot of times there's a lot.

A

There's many factors going on at once. So you probably know Asian is really really really really big and so I went I used to work in politics and I went to this session and they were talking about asian-american voters and trying to infer some like thing. That would be the same so that they could all figure out a strategy to get asian-american voters out of their house to vote for their candidates, um and that's just it's too big.

A

So uh we have like first-generation immigrants. We have second-generation immigrants. There are like at least a dozen different countries that people will come from and there's a huge different like difference in wealth, so this is from the Brookings Institute. So we talked about how there's this stereotype that Asian Americans are good at math um and that's true when, like other people from other groups, they have access to a good school. That's not rocket science. You live in a nice neighborhood.

A

You have a good school you're, more likely to be good at school, um but like that's not the case for all groups, and that has to do with when you came to the u.s. what your parents do for a living for a while for certain groups, you could only come if your parents had like an advanced degree like where a scientist or a doctor, or something like that.

A

It turns out that children of scientists and doctors of any race actually are pretty invested in school, and their parents are pretty excited to make sure that they get to live in a neighborhood where they go to a good school. So like trying to make this broad generalization of one group that may be a small part of your district based on like what, when you break it out, is like dozens of different factors, it's kind of ridiculous. So it's a really ridiculous session like um anyway.

A

So if you make a category that is like two generals like be careful of inferring like oh well, if that to people. So obviously we can make this huge inference like everyone from Iowa loves, Python and other things. You I wanted to stab yourself in the face of it, but.

C

A

Like you Tina, you can't extrapolate from too small of a sample, and you may find that, like a group that you're looking at is actually like, wait, teach it off, so I live in Boston, which has lobsters and bike lanes, telling me that nice picture so that we could enjoy them both at the same time.

A

So it's still like about 20 years ago, lessons started putting in more bike lanes, right and um and people started, asking like well cool, so we've been printing. All these bike lanes like are they working are more people, cycling are more people like using them, is it safer and they kept asking and, and in particular like car drivers were like, oh like it's like, we lost the lane over here like? Is it all worth it, and this city like stalled for a little while?

A

Finally, someone kind of leaked like we haven't been counting cyclists I was like wait. What and it's like nope we've been turning car drivers and literally everyone else. Strollers wheelchairs, cyclists, pedestrians, children, maybe squirrels. Who knows so in asking this question of like oh, so how are the bike lanes working they're?

A

Like we have them so they had said they do now count cyclists, separate from pedestrians and other users of the road, but they couldn't figure out like they were not able to say if they were having more or or less like traffic accidents if it was making cyclists safer or if anyone was even using the bike lanes because they weren't counting it so like you can't fix it.

A

If you can't see it so it becomes this thing of like well, you know we can't fix a problem that we can't quantify, and so you want to be careful. Putting like two things into one category, like my other car, is at Inchon. Novel I guess is sort of like in the spirit of Fenton, because he's little like kind of cerebral, but, like you know, you can't drive a book to work so um not so far as I've noticed so um yeah.

A

So putting things into the in, like you know, if you're, not counting it or you're, putting like too many things into one bucket you're, not gonna, have very useful information. So you want to make sure that you're not like taking the number of samples received and then using that to meet your categories and, like I said you want to step back so highly kind of like stretch our brains. So we talked about like we like stereotyped ourselves and we're.

A

Certainly when we were places that we know it's impossible, we're at so do we have you know some challenges right and in we have like a recent sexist homophobic society that we live in and have been like kind of soaking in for our entire lives, not our fault, but you know that's what we have so um so I would say.

A

The first thing is to educate yourself um and occur to try and step out of like what it's like to you and maybe think about like what it's like to be someone else right, like are you um but like so one of the things that I try to do in my work as a an intern amazing, for that is to follow people who are not like me, especially activists and advocates for groups that I'm, not part of so and it's great, it's passive. It doesn't involve them doing any work and me being like.

A

Oh can you tell me like I I need to know if you know I'm being sensitive to this like? Can you do a bunch of unpaid stuff for me and help me fix myself, so you can just follow people that are not like you like for a disability activist and and especially if you find yourself like in a situation where, like maybe you're like oh I, am building a community pool or not pool and I. We don't know any black people, it's like.

A

Well, you should get on that and follow some folks and and try and get a better sense of what it's like to be someone who's, not you up to until you're gonna like if you you know, you're doing this for your own edification, okay, fine! But if your company is paying you to like figure stuff out, then they could also pay someone from a marginalized group for their advocacy and expertise.

A

Everyone likes money. That's where that's why it's supposed to me: I, don't know what happened with LibreOffice um but uh yeah. It's that stuff!

A

It's work and it's work that we should be doing and thinking about so um another thing like I, said: I really like to think about the perspective of stuff and to get like a little bit more on what I'm doing from outside of what I'm doing it's it's the same when you're coding to like you know like sometimes you're looking you're, looking you're looking and then you're like all right, I'm, gonna, just shut the computer down go, get dinner and then you come back and you're like oh I, was on the wrong branch.

A

The whole time. That's why nothing worked and nothing was connected, so it it is the same with other work like stepping back and getting some perspective is really good like so you can kind of like.

A

Oh, maybe I'll read something from a different field or maybe I'll take a look at like what one of our like competitors, or we're all cooperative competitors and open source like someone who's doing something similar to me, but a little bit different like how did they approach the problem and so I think that the perspective is a really good one and taking a step back from what you're doing you can.

A

Also, if you are lucky enough to have friends that, like to you, know, complain about work like you can ask them to like your about a problem that you're working on and it can be a trade it could be over beers. It doesn't have to be like a painful thing, but um you know the kernel I get some perspective because some you sometimes someone will ask you a question that you will never have thought of like where you're listening and or they're listening and they're like oh well.

A

Where did the number come from and then you're like I haven't thought about that in months like wow, okay like but it's a great question and now I have to explain it to someone who isn't in like steeped in my work, and it gives you a little bit more of perspective and then finally, I would say to accept suggestions with grace so like if someone says like.

A

Oh that questions really weird and leading or I was wondering why, like you, were like this question, you know this survey is for all contributors and then I went through 37 questions about coding and no one asked me about translation or setting up events, which is work, and no one asked about like all of these other aspects of the community like helping people in IRC like so. If it's really all contributors, then where are all those other questions about the other types of contributions? That's a real world of examples. I won't name names self.

A

So once you get these suggestions, you may have to trim up some stuff and fix your things, and but this is better the earlier you can do it. The better, like you know it's most of those things don't go away or if they do, it means like a whole group of people, saw your stuff. One of them said something and then everybody else is like yeah, not worth it and went away. So you don't want that.

A

So, when someone gives you a suggestion, like hey, it seems like you're trying to do outreach here to, like you, know, women, undergrads and CS, and some of the language the mustache contest, not working. You know so like just accept that gracefully not with so much work. You can tell your friend that you complain to work about that part, but you don't tell the person taking the time to say: hey I'm not actually invested in your work, but it could be better.

A

So that's like a free gift right like questions, our love and patches should be welcomed, so um I have credits and I would be willing to take one or two questions. Although I know it's a kind of an abstract conversation, so thanks very much I.

B

Love the example of the.

C

Bike lanes, like they weren't being counted or they're in an aggregate and then you're trying to bring it into individual metrics to work through a challenge like that where there was an overgeneralization and then like, how do you even start to peel back something that abstract and go through it methodically with people yeah.

A

Well, I mean I, guess, there's two things: either you can take a look and see if there is something else. That might be? Oh sorry. So how do you work there? The aggregate data collection problem so there's two things like one. It might be that, like you, have another marker in there that helps you easily separate stuff. So, like you know it's uh if you're trying to be like, oh, we were looking at like all women developers, but not separating by college students and professionals, but you were tracking age.

A

You might be able to figure it out. So that would be great and lucky for you. But if not, then you probably have to start with you data to look at that problem. So yeah yeah in the back or in the middle.

B

A

Pull so the question was: how do you push the conversation forward when your bosses ask for a success metric, that's too simplistic, yeah I would give them not too many more metrics than they asked for, but, like maybe two more to kind of say like especially if one is like so here's the thing you asked for here's like one that almost completely contradicts the simplest sickness of the other and metric, and then here's the one you should have asked for.

C

A

You can't, but if you get like, if someone says I want one number and you give them 40, it's not gonna happen.

A

So how do you measure whether something is ethical or not? I guess yeah with the yeah, it's I think that's a really good question and it's like always evolving I- think we're always learning a little bit more about that, especially like, as you collect more stuff, I guess, that's what I'm advocating really is making sure that you're constantly stepping back and checking in on your work and looking at the way it affects the people that you're, you know, collecting information from to make sure that that makes sense.

A

In my mind, I think that diversity and inclusivity is its own good I mean you know without tanking the project, but that it helps us build better projects. It helps us build software that is less likely to exploit people or exclude people, and so so I would say like thinking about the diversity question that is in in some sense that that helps you get to like a more ethical build.

A

Does that answer your question, and that is a big way. We can talk to all right, one more you have it.

C

A

All the way, oh.

B

A

Oh and you know it so the question is: how do you account for the double of.

A

Yeah, so the question again is: how do you go back if you didn't record whether you know the duel marker of women and people of color again I? Think if you haven't tracked it and you you might be able to do it through some case study. So I got some information on that. But if you didn't I mean that's, why it's important to attract these things is that if you didn't track it, then you only have like sort of a gloss of how it went down and what it was like and so yeah.

A

So people should be tracking stuff. That is important and I. Think that's really important, and it's and it's like I, said it's. It is sloppy to make too few categories and then try to make inferences. So thanks for your question all right, thank you very much.