DDOS Prevention, Data Science Ethics -- Talking Tech w/ Jigsaw, Cloudflare, and Laura Norén
If you want to silence an advocacy group or media outlet, a fairly easy way to do so it to take down its website. Two organizations, Cloudflare and Google are doing great work to help nonprofits and journalists, as well as elections officials, protect themselves against cyber attacks. In this episode we talk to Alissa Starzak from Cloudflare and George Conard from Google about their efforts. We also share how you can take advantage of their free services (Cloudflare's Project Galileo and Google Jigsaw's Project Shield). <br><br>After that we talk to Laura Noren, a data scientist who was recently with NYU teaching a class on data ethics. What do data scientists need to know create technology that is better for society as whole?<br><br>More on Cloudflare's Project Galileo: http://bit.ly/2LDeYbJ<br>More on Google Jigsaw's Project Shield: http://bit.ly/2O3KZLG<br>Course syllabus from Lauren's class: http://bit.ly/dataethicsnyu<br>NYU Data Science Community newsletter: https://cds.nyu.edu/newsletter/<br><br>More on Alissa: http://bit.ly/2LoWRcS <br>More on George: http://bit.ly/2LwjMCW<br>More on Laura: https://twitter.com/digitalFlaneuse<br>More on our host, Brian: bit.ly/cdtbrian<br><br>Attribution: sounds used from Psykophobia, Taira Komori, BenKoning, Zabuhailo, bloomypetal, guitarguy1985, bmusic92, and offthesky of freesound.org.
Top Keywords
- ddos 0.012
- galileo 0.009
- jigsaw 0.008
- project shield 0.008
- cloudflare 0.007
- shield 0.007
- attacks 0.006
- election 0.006
- ddos attack 0.006
- george 0.006
- project galileo 0.006
- well 0.006
Transcript
Welcome to Tech Talk. Bye.
CT. Tea.
Welcome to CDT's tech talk where we dish on tech and Internet policy while also explaining what these policies mean to our daily lives. I'm Brian Wasilowski and it's time to talk tech. If you want to silence an advocacy group or a media outlet, a fairly easy way to do so is to take down their website. Two organizations, CloudFlare and Google, are doing great work to help nonprofits and journalists, as well as election officials protect themselves against cyber attacks. We'll talk about their efforts and share how you can take advantage of their free services. After that, we talked to Laura Noreen, a data scientist who was recently with NYU teaching a class on data ethics what do data scientists need to know to create technology that is actually better for society as a whole CloudFlare and Google are both doing a ton to help nonprofits defend themselves against a range of online attacks. One of the most common attacks is a distributed denial of service attack or DDoS. And through Cloudflare's Galileo and Google Jigsaw's Project Shield, nonprofits receive free support to fend off would be attackers to keep their websites live. Joining me to talk about their efforts are Alyssa Starzak from Cloud Flare and George Conard from Google. Welcome, Alyssa and George. Thank you. It's great to have you here. So first, tell me, you know, for folks that don't know and, frankly, me, what a DDoS attack is. What exactly is it? How common are they? Make this understandable for us.
So I'll I'll I'll start with this one. So a a DDoS attack is a way to try to stop other people from accessing a website. And we can get into the technical details of it, but one of the ways that I like to describe DDoS attacks is by using an analogy. So, you know, imagine if you have kids, imagine that you have a child who's about to turn nine years old, and you're gonna have a birthday party for for them and maybe eight or 10 of their friends.
That sounds like my nightmare, George. Well, it's gonna get worse. Okay.
It it may it may imagine that that there's someone who doesn't like you, you know, not that they don't like your kid, everybody likes your kid, but Of course. Nobody doesn't like you finds out about this. And, you know, I'm based in in New York, and I'm sitting in Manhattan, so I kind of, you know, think about it in these terms. You know, imagine that they post flyers all over Manhattan saying on this night, on Thursday night at 07:00, there's gonna be free cake at your address, and they post your address on that. And so the day of the party comes, and your your child's friends are trying to get to the party, but they can't even get through the streets streets because the streets are completely clogged with traffic. The sidewalks are full of pedestrians trying to get to the free cake. And even if they make it to your house or your apartment, they can't get through the door, because it's it's just full of so many different people trying to get there. And then imagine even if they did get through the door that, well, the person serving the cake has had a complete meltdown and is lying under the table and there's no cake left. Nobody's actually throwing cake. That's essentially what a DDoS attack is doing to a website. It's sending so much Internet traffic toward a site that either it clogs the internet connection to that site so no one else can get to it, or it causes the server itself to just stop responding because it can't handle the the entire traffic load. Wow. And no one gets cake.
No. That's a great analogy. How common are these attacks? And and who are the kind of victims or targets of these? So one of the things we talk about a lot is actually that they're relatively easy. So, it's not that hard to launch a DDoS attack. So if you have someone who doesn't like you, it's not that hard to launch a DDoS attack. Yeah. So they are relatively common.
And I'm guessing there's a range of reasons that someone might not like you. I think that's certainly true. Yeah. And a lot of non profits certainly working on, you know, difficult issues or issues that are contentious. I would imagine, you know, are a target. Do you have any examples of kind of nonprofits that you've worked with who who have been targeted?
So I can talk a little bit about how we started our program. Sure. So, so our project, Galileo, actually started in 2014, after our CEO noticed, that an entity that was on our free plan we have a we have a range of different options from a business perspective. One of them is a free plan. One of the the company the entities that was on our free plan that got a significant DDoS attack was a, a journalism, site, a media site Mhmm. In Ukraine. And the concern, obviously, as George alluded to, when you have your child melting down into the table because no one's getting cake, the same thing if you have a media site that's site that's not active, people aren't getting news. Aren't getting the information. Yeah. That's important.
So tell me, let's this might be a good time to kind of pivot to your specific projects. You went to it a bit. So, Cloudflare's Galileo, Google's Project Shield. Tell me a bit about what I know they're similar but different. So maybe, George, why don't you go first, and then Alyssa tell me a bit more about, Galileo. So, George, tell me about Project Shield.
Sure. And and and I'll start with just very briefly kind of who we are at Jigsaw. Because that's, you know, part of the story. So Jigsaw is a, an organization within Alphabet, which is go Google's parent company. We're very closely connected with Google, of course. And and we're a team of engineers and product managers, researchers, policy experts looking for ways that we can use technology to make people safer. And the, you know, Project Shield is something that going back to 2012, we had an engineer who was, in a country that had an election in that year and and working with their electoral board and found out that they they were suffering from DDoS attacks. And as Alyssa alluded to, you know, launching DDoS attacks have gotten easier and easier over the years, and cheaper and cheaper too. So you don't have to be a technical person to launch one. You can go out on the Internet, and if you know where to look, you can find somebody who will launch an attack on your behalf for under $5.
Woah. Under $5?
Yeah. Don't get any ideas. Yeah. No.
I have no ideas. I love everyone.
So we we started experimenting with ways that we might be able to put Google's infrastructure and Google's defenses, which we've, of course, built over the years as as Google has been a target of DDoS attacks, to use, to protect people who wouldn't otherwise be able to defend themselves. Either they didn't have the technical or financial resources to be able to do that. And so Project Shield grew out of kind of some experiments that we did in that space, and has has grown to be what it is today, where we are a free service that will offer protection to journalists and news organizations, human rights organizations, and the elections and electoral, sites around the world, as well as we do some work in the in the political space depending on the country that we're in. We can talk a little bit more about that later. But, you know, the the your your question about kind of how, how this impacts people, I think Alyssa said it very well that, you know, so many times you have information that someone is publishing, you know, whether they're whether they're a rights organization or they're journalists or or whomever. And the moment when they are ready to publish something that might be controversial or that someone might not like, is probably the moment that their readers need them the most. And so making sure that they're able to stay up and stay up under attack even if someone's trying to silence that voice, we think is very important. We have a mandate from the from the company to look at ways to defend free expression and and keep people safe. And so it fits fits very well into into kind of what we're doing. And you'd you'd asked earlier as well about kind of why, I I think we started doing this. And the, you know, the truth is that for for us, it's a two part problem. One is that core mandate that we have to use tech to to keep people safer. That's just what our job is at Jigsaw. But beyond that, we think and and I've had lots of conversations with others in the industry and including Alyssa and others at CloudFlare that, you know, DDoS is just bad for the Internet. And that makes it bad for business and it makes it bad for everybody. So the more that we can all do to work together to to stop these kinds of attacks, I think the better we all are.
That's great. So Alyssa, I'd love to hear even more about Galileo and what you're doing at CloudFlare. Yeah. I I wanna just say I agree with everything George said on exactly those issues. You know, I Cloudflare was launched in 2011, along some of those same lines. So our mission as a company is to help build a better Internet. And we actually launched in part with a free service, because what we saw at the time was, was different types of service for different kinds of entities. So big businesses that could pay for a lot of, different services, were protected from DDoS attacks. They had fast service. That was not true for small blogs, for small Internet sites. And and our sense was that that wasn't right from an Internet perspective. And Project Galileo actually grew out of those same ideals. So the notion that, somebody could be wiped off the Internet just because they couldn't afford to pay, particularly when those voices were particularly valuable. So voices that are talking about human rights, that are talking about, that are reporting on things from a from a, that are nonprofits that are that are helping people. All of those things are really important voices and our sense was that that is one of the things that the Internet has done. It's empowered those types of voices, and it was really important to make sure that that stayed the case. So in the reason we launched in 2014 as I mentioned was about this Ukrainian news site or that's that was the origin of it. But at the time, we weren't able to protect on a on a free plan everyone from every every size DDoS attack. So we could protect some DDoS attack at the time. It it wasn't across the board for our free plan. We've actually changed our our plan over time as we've gotten bigger. Now we actually provide free service, DDoS, protection services. So for everybody who signs up for our free plan. Project Galileo adds something onto that too though. It protects from additional cyber attacks. So Okay. We have a web application firewall that goes into effect for our our people who are part of Project Galileo, which helps them protect against, some common vulnerabilities on the
posting to our website on different things. What are you posting? Yeah. It's usually when I promote and then we have a great person who's like this this is for our own good that you're having trouble posting. So it all gets resolved and it's beautiful at the end of it. So those are fantastic projects. George, you also mentioned a bit that you do some work on elections. And I know that you've worked in, all of a sudden, you know, also kind of, you know, contentious countries where there's democracy may be at risk. What are some of the things you do around, like, elections and information and election security?
So the the as I as I said before, the, you know, kind of the origin of of SHIELD actually was was in talking to an electoral commission, in a in a different country. And so it's something that's been on our radar for for quite some time. And we see a lot of times, you know, digital attacks can correlate to conflict in the real world. And as we see elections become more and more contentious in different places around the world, including here in The US, and we see the capability of different actors, whether they're individuals, you know, all the way up to governments have more capacity to to launch digital attacks. You know, that correlation is just strengthening over time. And so we, you know, we've been working to protect electoral commissions, elections monitoring sites, and others like that since the since the beginning of Shield. And we've done that in a number of different countries around different elections. We we we've done two things in the last year to kind of go a little bit further on that front. So one is that last year, Jigsaw, launched something called Protect Your Election. And that's a so it's a website, and we can we can get the link for this, you know, out to to your listeners around in the Sure. On the website when it gets published. Protect Your Election really brings together a number of different tools that both Google and Jigsaw have built over the years to help people be safer during times of election. So that includes things like Project Shield. It includes things like two factor authentication to help protect, you know, your accounts and your email accounts and things like that, as well as for those that needed Google's Advanced Protection Program, which is kind of Google's highest level of security for individual accounts for people who are, you know, maybe highly likely to be to be targeted. And so we've we've brought all that together and have done a lot of work to help educate people about what those different threats are. From a Project Shield perspective, one of the things that we've seen is that during the time of an election, you know, the information that's important to voters is partially the elections boards or or electoral commissions or or whoever is running the election. Where do I go to vote? When is the vote happening? What's on the ballot? Things like that. But also what you know, the information from campaigns is is important too. You wanna be able to understand, the point of view of the different candidates to make an informed decision. And we've seen certainly in The US both, over the last several months and even just a couple earlier this week were reported DDoS attacks targeting county election boards, you know, are often much smaller. It may not have as many technical resources as well as even small local campaigns. Those were the two that were were announced that or reported on this week. And so we've we've we've started also offering protection to political organizations, not just the elections boards themselves or election election monitoring sites that campaigns, candidates, committees, things along those lines. We're doing that on a country by country basis. So we announced support for that in The US a couple of months ago. And then we'll be rolling it out to other countries as we navigate the various kind of local laws and regulations about what's possible and what's not. Cool.
Plissa, did you wanna share anything about Cloudflare's work? So we're we also launched a project called, the Athenian Project, that's specifically focused on US elections, and it's specifically related to state and local governments. You know, our sense was that what we were talking about right now in The US is really about the integrity of elections and George alluded to this too. You know, one of the big concerns right now is we people don't we want make we wanna make sure that people trust that the the the election itself is fair and that that the reality of an impact of a DDoS attack, if a website goes down on the night that that that things are being reported, that results are being reported, there's a lot of concern that people aren't gonna trust what's actually happening. Even if everything is absolutely fair, it's gonna look questionable and and we Because we can prevent that, because it's our core business to prevent that, our sense was that we should make sure that the entities, the the government entities that didn't have as many resources should actually have the resources to do it. That that was a really important thing. So I I would echo George's comments on that. It's it's incredibly important that people have faith in their in their systems and some of that is making sure that websites are protected. You know, I think one of the things that often happen, happens in the election space is that people focus on the election results specific or on the election voting specifically, the voting machines, any vulnerabilities there and they don't think about the perception issues. And I think that's a space that that that we can play in and we can play quietly behind the scenes that that Jigsaw can play in as well. And we can just help reinforce,
that these are things that just work. Yeah. Well, you all are doing amazing work. How can either a non profit, a journalist, an election election official get in touch with any of you? What's the best way if they wanna take advantage of Jigsaw or, Cloudflare's great services?
Go ahead, George. So for yeah. So so for for Jigsaw and Project Shield, Project Shield itself, you can find at, g.co/shield, or go to jigsaw.google.com, which has information on all the different projects that we have at Jigsaw and and links from there to to Project Shield. And the Shield site has a bunch of information about who's eligible, how the service works, so linked off to our health center to to try to help people understand what they're doing as well as how to get in touch with us, more directly.
And Alyssa, how about Project Galileo? So we have a couple different ways that you can get in touch with us. You can apply directly on our website to Project Galileo, which is just cloudflare.com/galaleo. We also, the way we actually work under project Galileo, because we're an infrastructure company and we don't deal we don't wanna be in the business of assessing whether particular organizations are worthy of our services or not, we actually partner with a bunch of very respected nonprofits like the Center for Democracy and Technology Thank you. To actually bring people our way. And so if if people are if they think are in need of services, we will protect them under Project Galileo. And what we actually typically do, when we get a request on our website, we will actually ask any of our sponsors if they're willing to to sponsor it or partners or if they're willing to sponsor sponsor the, the organization.
So there are multiple ways. Wonderful. Well, thank you so much. We'll make sure all of that is on the CDT website when we post about this one, and also every place that you can find Tech Talk. Alyssa and George, thank you so much for joining. It was such a pleasure. Thank you.
Thanks. It was great.
Build it first and ask for forgiveness later. The tech mantra, or at least it has been, but that might be changing. As data and technology become increasingly integrated into every aspect of our lives, the potential real harms to each of us becomes more apparent, whether we are thinking about connected cars or automated decisions on employment. This has led to ethics coming up in far more conversations about data and tech. And as our guest today says, you can patch software, but you can't really patch a person. Laura Noreen was recently a postdoctoral fellow at the Center for Data Science at New York University and taught a class on data science ethics at NYU. She joins us to talk about that course and why ethics matter in tech. Welcome, Laura. Thank you so much. So let's start very top level. How important is it to factor ethics into techno into the technology we create, especially when when it's a data driven technology?
I think it's absolutely critically important. And I think one of the big differences about, the the data saturated world that we're living in is that we have the capacity now to represent an entire human life in more exquisite detail and more completeness than we've ever had in the past. And this is something that we haven't, you know, the conversation is typically not framed in that way. Mhmm. But when you realize the the depth that we can that we can understand humans to, there's a huge amount of potential there, a potential benefit, social benefit, individual benefit, and then there's also kind of a great big dark side of things that can go wrong. And then things that go wrong typically don't go wrong for everyone. They go wrong for relatively small groups of people who, we as technologists really have a responsibility to look out for, you know, the the the 1% or 2% or 5% of of our populations that aren't quite fitting the model. So I think that's that's where I like to start the conversation with with technologists in particular because they're so used to thinking about things in terms
of an accurate, you know, let's move on with it or this is 95% accurate. We'll get to run. Yeah. We need to think about that that that extra 1% or 5% or whatever it is, that small group. And when we're talking about the 1%, you know, here or the 5%, we're certainly not not talking about the elite 1% typically. What who kind of tends to make up the groups that are, you know, affected more so by, you know, false assumptions or things being 95% correct?
Unsurprisingly, it's typically groups that have been historically underrepresented in public life. So whatever that you know, it varies from population to population. But, one of the examples that that is particularly illuminating that I like to use is the the example of self driving cars. Probably, self driving cars are a net benefit to society because it turns out that humans are at least Americans are pretty bad drivers.
I've been other places too, and there's some bad drivers worldwide.
Yeah. We we, we're not very good at it. And so we have lots of fatalities, forty thousand plus per year in The United States, and that's increasing even though we continue to build cars with more and more kind of safety features. We just get more and more distracted or I'm not exactly sure what the cause of our poor driving is, but it doesn't really matter. It exists. We're bad at it, and we're not getting any better at it. So probably self driving cars are in that benefit. That's that's actually a good thing. We should pursue this technology. But the the question becomes, well, when do we start putting these cars on the road? And the problem that that Uber just had and also the problem that Tesla had earlier, both of those companies were involved in accidents that proved fatal,
either for a driver in the case of Tesla or for a
pedestrian in the case of Uber, were obvious, types of problems that you could have predicted you might have with image recognition. Mhmm. One of them was, basically, an overexposure problem. There was too much light coming into the sensors, and they really couldn't figure out that there was a semi truck. Uh-huh. It seems pretty obvious that that would be there, but, it was a white semi truck, and the angle of the sun just made it very difficult for the sensors to detect. So the car ran straight into the semi truck. Yes. So that was the Tesla case. And that was several years ago. I think it's fallen out of public discourse. But that's how that happened, and that's a pretty predictable image recognition problem. And that's a 99%, 1% issue. When your technology is 95% accurate, it's gonna see that semi truck 95% of the time, but then 5% of the time, it doesn't. And that's, you know, that that to me is Doesn't seem like great odds. Yeah. Threshold. I mean, I don't I don't know what the exact percentages are, but that's typically how, you know, data science works. Yep. You're never gonna get something that's a 100% accurate. And so your question then is, where do we draw the threshold where it's safe enough to put this thing out put this technology out on the road? Typically, what we do in data science is we just compare to, well, is it better than what the baseline is now? If we put these self driving cars on the road, are they gonna be safer than typical humans driving? Right. And, of course, humans probably aren't necessarily gonna make that particular mistake driving straight into a semi. Although, there was a man in the car at the time, and he didn't react at all. Apparently, he was reading Harry Potter, which is very gross.
Depressing. It is. Absolutely. It's very encouraging.
So, but in the Uber case, they had kind of an equally predictable but other end of the spectrum problem with it. They their sensors were unable to detect this pedestrian in very low light conditions. Oh, that's not true. The sensors detected the pedestrian in low light conditions. They just couldn't tell what it was that they detected, so they decided it wasn't really worth slowing down for. Oh, wow. Okay. Yeah. So that's an example. Typical, like, you know, a lot of times we might look at image recognition and say, well, it's it's particularly bad at detecting, features in darker skinned people. That's true. That is and that's been true for a long time. If you look back at the history of film photography, we didn't have films that were capable, like, the the actual film itself that was capable of getting good white balance simultaneously for light skinned people and for dark skinned people in the same, you know, in the same image. Mhmm. So if you are a very white skinned person, light skinned person who has very dark skinned friends, you would almost never be able to get a good photo of both of you at the same time. So there's been some ethnographic research on, well, how how did the early not not today's, but pre digital, you know, camera people deal with this, and they just they wouldn't they wouldn't ever They wouldn't photograph people together? Oh my goodness. They would not do that. Or so, like, if, you know, David Letterman interviews Whoopi Goldberg or something like that, they would have a whole camera setup and a whole white balance, Oh, wow. Going on for him and a whole completely different setup with different lighting for her. Oh, wow. I did not realize that. Yeah. We knew this was a problem, but we didn't force it back to Kodak and say, alright, Kodak. Please fix. Fujifilm actually paid a little bit more attention to it, because they had they had slightly different populations using their film. But when we but so so tech ethics isn't some, like, impossible difficult thing. We actually have a fairly deep history and a lot of these technologies suggest we're gonna like, where we're going to have problems.
Fascinating.
So let's talk We haven't always we're as technologists, we type we we aren't always the best at knowing our history.
So tell me a bit. Let's go to that, you know, kind of the course that you were teaching when you were with NYU, and you've now moved on to Obsidian Security, and we'll get to that in a bit. But at NYU, the the course on data science ethics, what did your syllabus include? You know, what were some of the recommended readings that, perhaps I would like to read?
Sure. Well, we started with, moral philosophy. We do have, you know, some really deep theoretical precepts out there that kind of help to help understand well, you know, what are the appropriate trade offs. If you if you know your technology is never going to be perfect, does that mean you're just paralyzed and can't actually do anything? Because you can never achieve perfection, therefore, you can never release it. No. That's probably not gonna provide the the most benefit to society. But how do you figure out, you know, what kinds of technologies we should pursue when when it's appropriate to kind of start testing them in places where they could cause harm or could cause benefit. Mhmm. And so we start with something like, fairness principles that John Rolfe talked about. And I think that's pretty important for technologists. So he he had this experiment. It's a thought experiment that he liked to run to to determine how do you set the rules that should govern a group of people. And he said, well, obviously, this is not possible to do. But if you could imagine, like a a proto human being, sort of like a baby who's never had any experience before, but even more proto than that, who sort of has no has no history.
Okay. Like a new Furby out of the box.
Yeah. It's capable of thinking like an adult but has no history, and they are and they are blind to their own circumstance. We know that even babies aren't born into a blank slate. Right? They're born to particular kinds of parents who live in a particular city and in a particular, time and place. So we know that that already you're gonna come in with a set of, privileges and and affordances that other people may not have or vice versa. So but if you could imagine that you don't know where you're gonna be, you don't know who your parents are gonna be, you don't know which city you're talking about, which kind of, you know, what's the contemporary religion that dominates in this area. We you don't know anything, but you have to set rules. That's how you should think about, you should imagine that you could you could end up being, you know, the pedestrian who's out there on the street or the Uber Yeah. You know, the driver of this car or someone who is neither a pedestrian nor a driver, but just, you know, happens to be alive in society and is concerned about, I don't know, parking or, you know, other kinds of things that that cars deal with. And if you can come up with a solution that is fair for everyone but also is specifically not disadvantaging people who are already the most disadvantaged. So in this case, you would say, alright. We have to reorient our thinking to be most concerned about pedestrians who are not very good at, obeying traffic signals. They're probably the most disadvantaged. Right? They're most likely to have problems. So we need to make sure that our system doesn't disadvantage them even more than they're already disadvantaged. So I would say Uber failed that test because they weren't really able to look out for that woman who was crossing the street outside of an intersection in a in a poorly lit, you know, time and place. So they failed that test. But so we start with things like that because they actually, tend to be applicable across a whole bunch of different kinds of technological settings. Mhmm. And the thing about tech is it's innovative. It's always doing something different. That's the definition of what these technological fields do. So regulations are probably never gonna be all that helpful, in figuring out how you should build the next new thing. They're gonna provide some some bumpers, but they're not gonna tell you, like, when do you put this new technology out on the road? Because regulation is always reactionary. It comes after the fact. Yeah. So I try to I try to give my students some, you know, some capacity to have an ethical imagination so they could think about, what they should bring to the table. What kind of history do you need to have before you, you know, start developing technologies in a field that isn't really your domain. Your domain's computer science. You may not know that much about, like, the history of film photography. Why should you, frankly? But you might be able to start asking yourself questions that lead you to look into those kinds of domains or collaborate with someone who knows better about those domains. So that's always it's always a strength. You'll learn a lot even if it turns out not to have that much to do with ethics. And you'll also get some idea about how to make decisions about what to build, when to release it, how much testing to do before you move on to the next, you know, part in the product development cycle. That's great. We also talked about, you know, like, how much we read an article that kinda stuck with me which is how much how much activity does does a person say Facebook account need to have before Facebook can start predicting things about them? Their gender, their race, their age, whether they're homosexual or heterosexual, and what kind of data is really the most telling. And that was a pretty interesting article because they really Facebook doesn't really need to have that much activity from any given user because their background, you know, their base their database of of all people is so huge that just a few things can sometimes be extremely telling. And a few activities might be something as silly as saying that you like a Britney Spears album. She's not exactly releasing new material right now. So
Oh, if only we were that lucky. But
So that's that's actually more like, if you if they predicted that you're say a man and that you're liking Britney and they can see that you're liking Britney Spears, that's more indicative of your, sexual orientation than whether or not you, I don't know, changed your profile picture to a rainbow flag after DOMA. That's interesting. Because lots and lots of people did that. So it turned out the political statements about sexual orientation were not very revealing. But the kind of the cultural stuff, the things that people are like, oh, I don't really care if Facebook knows that I like Britney Spears or that I, you know, commented on this picture from my local coffee store. It's actually those things turn out to be Oh, that's fascinating. Really healing.
That's fascinating. Who were the type of folks that that took the course? Was it a a cross of, you know, data scientists, but also people from different disciplines? Or who who who was there? Because it seems like a lot of people should be taking this.
Absolutely. I think so. But in fact, it was it was all of my students were extremely courageous because this was not a required course at the time. An elective. It was an elective. And so this was this was a group of people who were were capable of saying to themselves, gee, I maybe don't know everything about the world and I'd like to be able to, you know, move sensitively through this world and build things that don't hurt other people or that, you know, bring the best benefit. So it was a range of people who come from more of a liberal arts background, media studies folks, sociology people. We had someone there who really wanted to take the course but couldn't because it didn't fit her requirements. So she came out of, like a nutritional counseling background. How interesting. You know, data in fact is saturating so many different CEOs that really the the the population for this course could touch on pretty much anyone who's planning on having a career. But and then there was also kind of a core group of people who are data scientists and who are aware, of the power that they have and aren't quite sure how to assess whether they're using it for good. They care. Just because you care doesn't mean it's always straightforward what what
should be done next. So now we're seeing a bit of, you know, for lack of better word, tech lash. It's a term that our president's starting to use and other folks are. So kind of anti tech a bit. Do you think ethics and now that you're kind of in the private sector with Obsidian Security, you probably have a different perspective as well and don't have to comment on them. But do you think that ethics and kind of, you know, this upfront thinking through ethics might be a way to turn the tide back towards, you know, pro innovation and pro tech?
Well, if I were rebranding things, I'd probably how they how they'd like to impact humans, all humans, not just sort of I feel like a lot of the apps that we have these days are basically about instant and complete gratification. Let me push a button and a car shows up to bring me somewhere. Let me push a button and I get all of the music I like and none of the music I don't like. It's it's a it's an let me get, you know, DoorDash or Seamless to show up with the food I like quickly. So we've we've created kind of we've we've been through an instant gratification moment, and I'm hoping that the next moment is a little bit broader than that because those technologies really don't benefit everyone. Yeah. They benefit the button pusher a lot more than the person who's driving the car or delivering the food or, the artist who make those produce those songs. Yeah. So we've kind of before we we've tipped we've tipped in a direction that does really really require some deeper thinking about what kinds of technologies we wanna build or what kinds of society we want to live in. Well, I hope all your students or your former students are gonna do that.
We're almost out of time. But before I let you go, I'd love to hear a little bit more about kind of your prior work before you got into, this area. You are actually the co editor of a book, toilet, public restrooms and the politics of sharing, and you've done a lot of research on workspace and technology. Can you give us, like, a quick, you know, thirty seconds or one minute, you know, what you what you've been doing there? Because I think it's fascinating.
Well, it turns out the politics of sharing really applies to data sharing too. Mhmm. The the the problems that people have in the bathroom, they're essentially very private data that's out in semi public spaces and the the tensions and the concerns are very, very similar. It's about Fascinating. Like, am I gonna get germs? Am I gonna get diseases, viruses? What what to do about the other we just share spaces and even and what to do about our own dirtiness, if I can. Messiness.
No. Those things are Our online messiness. Google. Yeah. A lot of cleaning up to do. Well, Laura, it has been a joy chatting with you. I've learned so much. Thank you for joining Tech Talk. We'll have to have you on again. And will you give us a few links that we can put with the podcast for our listeners to, to learn from your course? Sure. I can send you the whole syllabus if you wish. Alright. Awesome. Well, we will put those on our Tech Talk blog post and wherever we post it, which our producer knows more so than I do. Laura, thanks so much for joining.
Thank you so much.
That's it for this episode of Tech Talk. Be sure to take a look at Laura's syllabus for her data ethics class, which is available in the blog post for this podcast at cdt.org you'll also find the links there for the free services from Cloudflare and Google I'm Brian Wislowski thanks so much for listening
