Technosolutionism, Ethical Technologists And Practical Data Privacy
Subscribe to:
double up
Shane Hastie Good evening guys. This is Shane Hastie for the InfoQ Engineering Culture podcast. Today I'm sitting with Katherine Jarmul. Welcome Katherine. Thank you for taking the time to speak with us.
Ekaterina Yarmul Thanks so much for the invite, Shane. I am very happy to be here.
Introduction [00:34]
Shane Hastie We met last year at QCon in San Francisco, where you gave a challenging and interesting presentation on technology solutions. But before we get into that, it's worth asking "Who is Katherine?" to start asking
Katerina Yarmul Who is Katherine? I am currently the Chief Data Officer at ThoughtWorks and recently published a book called Practical Data Privacy. I see myself as a privacy advocate, privacy engineer, and machine learning engineer. I have been in technology for a long time and am interested in the intersection between the political and technological worlds.
I also founded PyLadies, the original California chapter in 2010 and 2011. So for the last 10 years I've been working a bit in technology, machine learning, machine learning and data science, about five years in particular. And we see how this progress is linked to the problem of technoanalysis.
Shane Hastie Perhaps a good start to this conversation would be the question of what is technoanalytics?
Definition of technoanalysis [01:24]
Ekaterina Yarmul Yes, I think I described it in the discussion, and I describe it in the same way that we have a magic box of technology and we take the problems of the world, or the problems created by other technological boxes, and their problems. : and store it in a magical tech box. And then comes happiness, solving problems, everything is fine.
And I think you can put that abstract history into what I would call technoanalysis. So the idea is that if we have another technology, we will solve this problem.
Shane Hastie I remember I was making a process model, and you put the cloud in the middle, and the cloud has the letters ATAMO, so a miracle happens, and after coding something happens. So now we are replacing ATAMO with Cloud, then technology comes and things get better. But why not?
Katerina Yarmul Why doesn't technology solve our problems? I think one of the things you have to think about when looking at the history of technology is that it's largely driven by innovation and the desire to create and change something, simplify something, or solve a problem. something like that Certain human desires, such as the desire to die, the desire to destroy, if we look at the history of technology...
I think I linked it to the discovery of gunpowder in my talk, and when they discovered gunpowder, they were actually trying to find the miracle of life. They're trying to find a magic potion that will solve people's problems, and they've discovered and created gunpowder. So if you think it's a beautiful metaphor in our minds, I'm not anti-technology, I'm pro-technology. I work in the field of machine learning and privacy.
But we also have this human form of man-made technology that is a reflection of what we see in the world. And when we do that, we're essentially hiding our biases, our expectations, and our ideas about how to solve a problem in technology. So we cannot separate what we think the solution should be and how we create the technology.
And I think one person's solution is another person's problem. And I think there's probably no universal morality or universal truth, so it becomes a difficult subject when you think about it. I'll build something that works for me and then expand that to work for everyone else. And where does that take us? Because depending on the context, the results can change.
Shane Hastie We do not investigate unintended consequences. Our optimists, optimists... How do we technologists take this step back and not consider the unintended consequences?
It is necessary to identify and analyze the unwanted effects [04:56]
Ekaterina Yarmul Yes, that's a very good question. Again, I don't think there is an answer. I think one of the things we need to think about is how we think about the history of our industry and technology. This is what haunted me for years. We constantly think that we are inventing something new. And if you really look at the history of computers and computing, and you go back to the history of mathematics, the history of science, you start to see patterns, and you start to see those patterns repeat themselves over time.
And so I think a useful starting point for most people would be to look at the history of your industry, whatever technology you're in, and the history of your industry. If you're in a consumer-oriented industry, maybe you're looking at the history of consumerism and things like that, and I think what the big issues are today, what's been created and what people have wanted to do. : What are the solutions and workarounds for this?
Applying that curiosity and maybe exploring that curiosity in something before, assuming I'm the first to come up with this idea, I'm the first to have this problem, and I'll be the first to solve it, who knows, it sounds too naive, but I think I should be there.
Of course, I was there when I found the problem and wanted to help fix it. And I think it's a very exciting and exciting story, what technologists have told us a lot is the awakening of the age of Silicon Valley and we're going to innovate and do things differently.
And I think it's good to have an optimistic energy. I am a Californian, very optimistic. I often root for the team. I have such strength, it is in the culture. But I think we can also use curiosity and humility and take a step back and look at past experiences to better understand how we can exceed our investment expectations, our capabilities, and our technology capabilities. General solution to the problem.
Shane Hastie But we don't teach budding technologists to think like that.
Expanding our horizons through participation in various disciplines [07:40]
Ekaterina Yarmul I know: it's really interesting. So one of the things that I've talked about in the discussion and I think you and I have talked about a little bit is why we don't have interdisciplinary teams. Why don't we have a group with a historian, ethicist or philosopher, where there is community group participation, so that communities can participate "equally" in solving problems?
And I think my experience is a great example of a lot of people who have worked in ethics and technology, and I'm not a specialist in the discipline. I went to school on a scholarship to study computer science because I was really good at math and I loved math, but I went to artificial intelligence school in the winter and I really hated Java.
We mostly worked in Java and I was really bad at it, so I switched to political science and economics, I could still study statistics and statistical logic, which I loved, but I didn't have to deal with boring Java applets. : what I didn't like very much.
And so I think there are people I meet along the way who have careers like mine, or end up in the ethics and data or ethics and IT space. I think a lot of these people study other subjects at the same time and then come back to technology, or start doing technology, go somewhere else for a while and then come back.
And I think that we, as an education system, need to think and reflect on the ethics of technology from the very beginning, even in elementary school. After all, it is the ability to think and learn. It's not like you learn everything in one day and learn by yourself. This is not what you should do in the context of a university.
Shane Hastie Bringing that curiosity. Before we started recording, we had a little fun: what do we hear in an echo chamber and how do we know which one it is? So if I really want to know how to get out of my echo chamber.
Get Out of Your Echo Chamber [10:14]
Katherine Yarmul : It's very difficult. Now we have algorithmic merging systems that ask you to stay in your cell or move to an adjacent cell. And I think it's difficult. I don't know what your experience was, but I think that especially during Corona and before people didn't travel, it was very difficult to figure out how to communicate with people from different fields and disciplines. Apparently, some of them are starting to disappear.
So the conference starts again. We met and talked at a conference. I think that's one way, but another way is to specifically ask yourself if you want to get out of your psychological comfort zone and I don't want to put anyone at risk, so obviously if you want to, that's within you. to arrive Are you afraid to learn about it, or are you excited or resistant to research?
And I think sometimes there's a bit of excitement when you think about it but you always find an excuse not to, maybe it's a way to get out of where they're stuck and find new ways. . And that thinking has a lot to do with human psychology and a lot of thinking about communication and community.
So these are not my ideas, they are just existing ideas. And I don't know, I'd be very interested to know how you get the filter bubble out.
Shane Hastie Personally, I go out to meet people I don't normally meet.
Ekaterina Yarmul How do you fit into the new environment?
Shane Hastie I will go to a new environment and introduce myself with curiosity.
Ekaterina Yarmul Wonderful.
Shane Hastie I don't always get it right.
Ekaterina Yarmul I think that's part of it, knowing that sometimes they're going to be uncomfortable or they're not going to go the way you want them to, right?
Shane Hastie : Yes. I had a wonderful experience. In my day job, I usually teach IT and business related subjects and have had the opportunity to teach a group of people who teach nursing, plumbing, health and hairdressing.
Ekaterina Yarmul Wonderful.
Shane Hastie And it was a completely different group. What they had in common was that they were all PTV teachers and I was a PTV teacher. So the teaching aspect was general, but their audience, the subject matter, the subjects they teach 18, 19, 20-year-olds who are starting their careers, as opposed to maybe working with people who are mid-career. , is informative. it was three days.
Ekaterina Yarmul Yeah, I mean sometimes, it's just… I have a few activist groups that I work with that are people from different backgrounds and backgrounds, and I think sometimes I feel like having these kinds of conversations. I've noticed lately that when I've been away and kind of in my tech bubble or my friends' normal life bubble, it can be really refreshing to come out of the same topics over and over again.
Shane Hastie Keep your new book a secret. Tell us a little bit.
practical information about privacy [13:51]
Ekaterina Jarmuil I wrote this book thinking it was the book I wanted when I first became interested in privacy. And first of all, I was interested in thinking about ethical machine learning. So how to make machine learning more ethical and inclusive, and how to deal with stereotypes and prejudices, the social biases we see when building large-scale models, is an important question. today
But, as I mentioned in the conversation, when I was looking at my own technological solutions, I thought that there is nothing I can do in a technical sense to correct the social biases that appear in these models. And the researchers are working in this place, I appreciate their work. And I think when I evaluate, I think, I can make a meaningful contribution here, and I think that contribution will really contribute to the field in a meaningful way and therefore achieve the goal of my work?
After all, this is not an answer I made up. And, of course, this can be a difficult time, but when I was interested in privacy as well, I had hope. And I saw that privacy has a lot to do with thinking about the ethical considerations of using data, because of the concept of consent, should we use that data? Can we use this data? Should we ask people if we can use their data? It was very interesting for me.
And the more I got into the secret, the more interesting it became, because there's a lot of great math. So it's been this combination of... Well, that's what I think I can contribute. Two things I love: math and maybe helping society in some way, no... technology won't solve everything, but it's a positive contribution I can make to the world as a technologist.
and when I first got into this field, it was professors and especially graduate students who had spent years studying, say, cryptography or differential privacy or these high-tech concepts. And while I pride myself on my ability to read scientific research, it's about thinking about those ideas and becoming someone who knows how machine learning works. Does anyone know how these privacy technologies work?
So when O'Reilly called me and asked if I'd like to write a book about privacy technology, I agreed. And I said that I would be very happy to target people like me, people who are experts in mathematics and data science, who have worked in this field, who have noticed that there are privacy problems that they want to solve and have. He heard these words, but still continued. தித்துர்யு does not have proper introduction then how to apply in real life தியப்பாஷ்கிக்கை ஜையையை.
So each chapter has a little something, we start with a small chapter and we learned some basic concepts. and then we have the Jupyter code and the notebooks that come with the book, well, there are open source libraries that you can borrow. Here's how you can use them. Here's how you can apply them in your technology context, whether it's data engineering, data science, or any other field in the world of machine learning or programming.
Shane Hastie Can we have your favorite? Which of these privacy technologies do you like best and how do you use them?
Practical application of data privacy - federal data analysis [17:32]
Katherine Yarmul yes One of the things that excites me most about changing the way we work is thinking about federated or distributed data analytics or federated or distributed learning. and the idea is -- there are already systems that do that, but the idea is that the data is really always in the hands of the user and we don't really collect the data and store it centrally. Instead, we may deliver machine learning to a personal device or provide machine learning. GPT4 Everyone is here, so that people can manage their own learning through self-learning.
And let's say that we run federated queries on some kind of device that analyzes data on something national, and we could run them too. And many times when we want to confirm these उटपधणप०णप०रा र, then we can also add अंजे चाई प रूयूज़ी. அம்பர்புர்பிர்ஸியில் பெர்கியு யாட்டை பர்ரிர்ப்பு ப்பு ப்புர்ப்பு பர்ரிரை চালনা করার অনুমত়ুমত়ুমত় 0 - রতেও সাহায্য করতে রেও সাহায্য করতে রো রেও এটি ডিক্রিপ্ট করা সুতরাং, এনক্রিপ্ট - রা ি, এনক্রিপ্ট তখন তখন আমরা ডিরা ডিক্রর্রর্র উ দাহরণস্বরূপ.
and all of this can run in a distributed sense, which will significantly increase the average privacy and privacy for people's data and will be, because we both probably recognize a fundamental change that I'm not sure will happen in my career, but I will want it to happen. I think it's going to be really cool, and we'll see how the wind goes.
One cool thing is the Google memo that was leaked recently that we don't have a mine and neither does OpenAI, specifically mentioning the idea of training their open source models for their own personal use. And if you fear Google, maybe it will become a reality.
Shane Hastie : What about the social aspects of the data privacy team? So, building cross-functional teams, dealing with privacy advocates, dealing with legal issues, etc., how do I communicate as a technologist?
The need and challenges of multidisciplinary teams for data privacy [20:05]
Catherine Jarmul : Yes. And I think we, again, this multi-stakeholder action is happening directly within an organization when we talk about privacy issues, we usually have at least three main stakeholder groups that all speak different languages. We have legal compliance or privacy representatives who also have some type of legal or regulatory understanding. We have information security or cybersecurity or whatever you call it in your organization, who have their own language and their own understanding of what privacy means and what security means.
and then our real technicians are implementing what needs to be implemented. and we have our own language, which is sometimes shared and understood by those other groups, but sometimes not depending on how we specialize. So especially when we look at specializations like machine learning, it can become quite difficult for a legal representative to argue about privacy leaks in machine learning systems because they haven't studied machine learning. and they may or may not know how the models may or may not store outside information or other private information as they are trained.
And so when we look at these areas, and if you ever want to get into the field of privacy engineering, you kind of bridge that conversation and you act as a spoke to allow these groups to share their concerns. identify risks, and to assess those risks, assess whether they will be mitigated or whether they will simply be documented and accepted and moved on. and I think that's why the field of privacy engineering is growing when we see things like this, I'm not sure if you've seen the meta fine announced this week of 1.3 billion euros, which meta was fined for transferring a bunch of personal data from the EU to the US and storing it in US infrastructure.
These things actually affect everyone. This is not just a special case. It's real control and real location of what's happening. And thinking through privacy in your architecture, in your design, in your software is increasingly expensive if you don't do it and increasingly, I think, important to people. I think beyond just thinking through regulatory aspects, I think there are exciting aspects to being able to tell your users, we do data differently, we collect things differently, and we can definitely start to see that marketing is starting to be pushed around. In particular, we offer something that is more personal than our competitors.
And I think that for better or worse, I'm maybe a little cruel. I don't necessarily think it's from the hearts of CEOs around the world. I think some of it might be a genuine consumer demand for privacy. And I think that people get scared when they find out that things are tracking them and they don't expect it. and I think that privacy by design is finally hitting its stride after almost 30 years of writing this, it's finally hitting the same thing that people want and what we as technologists should be thinking about implementing.
Шен Шести : Because these are things that can't be done, they can, but it's very, very difficult and expensive to recover them later. At the core of the systems you are designing and building, they need to be accurate.
Privacy needs to be at the core of architecture and design, refactoring for privacy is possible but very difficult [23:42]
Catherine Jarmul: Absolutely. I mean, I think there's more ways to approach it, it's your specialty, iterative and agile perspective. So you don't just rip out the core of your system and say, "Oh, we're going to refactor it for privacy. We'll see you in two years," but think, OK, that's why risk assessment is really helpful, especially multi-dimensional. , combine the group. and where is everyone's biggest fear that they're not talking about security breaches or privacy breaches or some other bad publicity and start prioritizing those and see if there's a small part of it that we can actually take and redesign? with privacy in mind?
অথবা এমনকি একটি স্বপ্নের অধিবেশন থাকার, যদি আমরা নকশা দ্বারা গোপনীয়তা করি তবে আমাদের স্থাপত্য কেমন দেখাবে? এবং সম্ভবত সেখানে এমন কিছু আছে যা আপনি বলতে পারেন, "ওহ, আমরা কিছুক্ষণের জন্য সিস্টেমটি প্রতিস্থাপনের বিষয়ে ভাবছি। আসুন এখানে শুরু করি।" এবং আমি মনে করি যে গোপনীয়তার ছোট অংশগুলি বাস্তবায়নের উপায় রয়েছে এবং আমি কখনই কাউকে বলতে চাই না, "ওহ, আপনাকে সবকিছু পুনরায় প্রয়োগ করতে হবে।" আমি মনে করি এটি অবাস্তব এবং শাস্তিমূলক যখন আদর্শটি নকশা দ্বারা ব্যক্তিগত জিনিস তৈরি না করা। আমি মনে করি আপনার নিজেকে অভিনন্দন জানানো উচিত এবং আপনি বর্তমানে যা করছেন তার চেয়ে ভাল কিছুর দিকে প্রতিটি ছোট পদক্ষেপে উত্তেজিত হওয়া উচিত।
শেন হেস্টি : ক্যাথরিন, এখানে কিছু সত্যিই আকর্ষণীয় এবং কিছুটা চ্যালেঞ্জিং বিষয়। লোকেরা যদি কথোপকথন চালিয়ে যেতে চায় তবে তারা আপনাকে কোথায় পাবে?
ক্যাথরিন জারমুল : হ্যাঁ, তাই স্পষ্টতই আপনি প্র্যাকটিক্যাল ডেটা প্রাইভেসি বইটি দেখতে পারেন। এটি শীঘ্রই শারীরিক আকারে পৌঁছানো উচিত, আশা করি, আপনার চারপাশের বইয়ের দোকানে বা আপনার প্রিয় বই খুচরা বিক্রেতার মাধ্যমে। কিন্তু আমি সম্ভবত প্রাইভেট নামে একটি নিউজলেটারও চালাই, এবং আমি অবশ্যই বলব যে এটি সুপার নর্ডি। তাই আমি সেখানে একটি সতর্কতা দিতে চাই। এটাকে সম্ভবত প্রাইভেট বলা হয়। এটি সম্ভবতprivate.com-এ, এবং এটি বিশেষত সম্ভাব্যতা, গণিত, পরিসংখ্যান, মেশিন লার্নিং এবং গোপনীয়তার মধ্যে এই সংযোগস্থলের চারপাশে, অবশ্যই, কিছু রাজনৈতিক মতামত এখন এবং তারপরে নিক্ষিপ্ত।
শেন হেস্টি : অসাধারণ। ওয়েল, আজ আমাদের সাথে কথা বলার জন্য অনেক ধন্যবাদ.
ক্যাথরিন জারমুল : ধন্যবাদ, শেন।
উল্লিখিত
. এই পৃষ্ঠা থেকে আপনি আমাদের রেকর্ড শো নোট অ্যাক্সেস করতে পারেন. তাদের সকলেরই ক্লিকযোগ্য লিঙ্ক রয়েছে যা আপনাকে সরাসরি অডিওর সেই অংশে নিয়ে যাবে।
