“Data is not neutral.”
— Sakthisree, in this episode, on the power relations that inform the extraction and manipulation of data in our lives.
Machine learning and artificial intelligence may have become buzzwords, but little scholarship exists on examining the sociological impact of the constant, ceaseless extraction of data from our lives and our very bodies. How can we look at technology through the lens of social justice? How can data impact us, and how can we incorporate feminist perspectives in how we deal with data? How do power relations inform the extraction and manipulation of data? In this episode, co-hosts Sanchi Mehra and Vandita Morarka are joined by Sakthisree, a young engineering graduate from Chennai and a Citizens for Public Leadership Fellow 2021, One Future Fellow ’20, Youth Panel Member for CMH-Global Mental Health Data Bank Project, Machine Learning Engineer, and Intersectional Feminist, as they discuss these questions, and many more.
***
Love to read or want to revisit your favourite bits? Dive into the full transcript below!
Vandita
Hi everyone, welcome back to the Nurturing Radical Kindness Podcast, a space where we explore radical kindness as a pathway to achieving social justice. My name is Vandita and my pronouns are she and her.
Sanchi
And I’m Sanchi, and my pronouns are she/her as well. Thank you, everyone, for joining us today. We have been loving your engagement on all our previous episodes and are super excited to have you listen in as we talk about data feminism today. Jab hum data feminism ki baat karte hai, (when we talk about data feminism), in essence, it means relooking at data in tech and relooking in a manner that is informed by an intersectional feminist lens. However, data feminism ke aur bhi bohot saare aspects hai (has a lot of other aspects) and we will be looking at some of these in our episode today. Toh shuru karte hai (let’s start), because we have a lot of interesting things to discuss. And to start us off, Vandita, why don’t you tell us a bit about what data feminism actually is? Yeh hota kya hai? (what is it about)
Vandita
Thanks for that question, Sanchi. It’s such a fundamental question to even start thinking about what this concept is and what it looks like. So in the book titled Data Feminism, the authors Catherine Dignacio and Laurel F Kline, mention data feminism as a way of thinking about data and its analysis and display that is informed by the tradition of feminist activism as well as feminist critical thought, which is to say that, you know, the way we currently collect data is fairly capitalistic, it’s extractive, it positions people as a product, and data feminism is a way to look at data and see what is it that we can do with data that is beneficial without creating any sort of negative harm and consequences. Data feminism also has very practical elements to it. It helps us relook at how inherent biases get crept up into data, how machine learning, artificial intelligence can become biased as well, and how over the long term some technological tools can often do more harm than good. Because they perpetuate and they create first, and then they perpetuate a new kind of structure of oppression that replicates the offline world. So, for example, thinking about who are we collecting data about, who collects this data, what does that say about the sort of data recording we’re doing? All of these are important questions when we think about data feminism and data feminism also examines data science and technology through a lens where you identify power and how it is often skewed, and then you work towards correcting that skewed power balance.
Data feminism teaches us to look behind the data to see who is recording this data, whose data is being recorded, analysing it, accounting for inherent biases and not just say, ‘Okay, this is the data’, ki ye data hai aapke saamne toh bas ise accept karlo (this is the data in front of us, so just accept it). Jaise, you know, hume koi bhi WhatsApp message aata hai hum use aise hi accept kar lete hai (when we get any WhatsApp message, we just accept it), but data feminism asks us to question it. It says ki agar apko koi message, jaise agar aap WhatsApp ke baare mein sochte hai, agar apko koi message aata hai, apko sochna chahiye ki voh message sahi hai, galat hai, sach hai, nahi hai (When we get any message on WhatsApp, we think about whether it is true or accurate). Same way, data feminism asks you to think ki aapke saamne jo data dikha rahe hai, voh sahi hai ya usme biases hai (if the data in front of you has any biases). Apko aise questions puchne chahiye data se aur voh data collection process se to understand ke uske piche koi biases hai ya nahi (we need to ask questions to understand if the data or data collection process have any biases behind them). And then it also goes on to look at how pressure can be interlinked, how digital spaces mirror on-ground situations and how this can become a way for a lot of people to hold on to privileges, right? In fact, you know, Sanchi, I’m gonna also ask you, because we’ve been having this conversation so much, why is data feminism so important? And why do we need to talk about it and how it impacts us?
Sanchi
Right. Thank you so much for that. And I think, Vandita, you’ve already laid the ground for why we need to talk about it and why it is important to us and we will be coming back to this with our guest for today. But to start us off, like you said, aaj (today) data is all around us, right? And what data feminism does is that its first step is to just examine power structures like you said, and it really sees, you know, how power operates in the world. And ek baar ye analysis hogaya tab (when the analysis is done then) it moves to see how it is that we can challenge these power structures. And of course, going back to what you said, it’s important to remember that data in itself is not objectively good or bad. It’s how it is used and how that serves only some people is what we have to take into consideration, and data feminism actually helps us do exactly that. Jo humara data hai (our data), it reflects existing power structures of course. And ek example lete hai yahan pe (let’s take an example here) which goes back again to what we were starting with, which is how it is that the data that we put into our systems can lead to machines mirroring them. So, the inherent biases and the circle that it perpetuates. So, agar jo humara data hai, usme hi biases hogi (if our data itself has biases) then the machines will mirror those biases. Say historically, if we have hired more men than people of other genders for senior managerial positions and hum ye data ek AI machine ko feed karenge (if we feed this data to an AI machine), then this bias will get replicated when the machine makes a decision. So, the algorithm will then automatically be biased in favour of men and ye cycle chalta jayega (the cycle will go on).
So, this is where data feminism steps up to help us, first of all recognize existing power structures, and then it helps us challenge them. It really just helps us rethink the binaries and hierarchies that we exist in and because we recognize them, it also pushes us towards accepting pluralism, because data feminism yeh kehta hai ki (says that), knowledge comes from lived experience and it tells us that koi bhi (any) complete knowledge agar kuch hai (is anything) that comes from a synthesis of multiple perspectives and usme bhi (in that also) it gives priority to local and indigenous ways of knowing. So, this means that knowledge stems from combined efforts rather than individual mastery. And iske saath hi (along with this), another very important thing that data feminism does is that it also makes the labour of these multiple hands visible. And jaisa ki hum feminist leadership principles mein bhi dekhte hai (like we see in feminist leadership principles also), jab success ya achievement ki baat hoti hai toh (when we talk about success/achievement then) these are not credited to just one leader. Sapka labour, sapka effort (everyone’s labour and effort) and again the knowledge that comes from the community is not only acknowledged but it’s also really valued. So, if we go on trusting data blindly, it could lead to perpetuating more harm, as we’ve already seen. Jaise ki USA ke kuch jagaon pe risk assessment algorithms use hote hai to inform prison sentence ya bail granting decisions. (In some places in the US, risk assessment algorithms are used to inform prison sentences or bail-granting decisions) Research has shown that through this algorithm, black defendants are mislabeled as high-risk more often than white defendants. And along the same lines, there is this brilliant mixed-media installation by Mimi Onuoha. It’s called the ‘Library of Missing Data Sets’, and in the artist’s own words, “it is a physical repository of those things that have been excluded in a society where so much is collected”. And they go on to say that what we ignore in our world reveals so much more than what we give our attention to. So, the piece is based on data feminist principles and what it does is that it’s helping us look at the existing power hierarchies through what is considered worthy enough to be collected or what’s worth paying attention to, and that’s only a bit about why data feminism is so important especially today. And I think now would be a perfect time to introduce our guest for the day and get to hear from a practitioner. Right, Vandita?
Vandita
Oh, definitely. I’m so excited about our guest for today because they were someone, I initially was able to externally start exploring my interest in data feminism with and I’m so glad that all of you get to listen to them as well. So, let’s introduce them, Sanchi. So today we have Sakthisree joining us. She’s a 21-year-old young engineering graduate from Chennai. She’s also a Citizen for Public Leadership Fellow, 2021, One Future Fellow for 2020, and a youth panel member for CMH Global Mental Health Databank Project. She’s also a machine learning engineer and an intersectional feminist. Sakthi is passionate about looking at data through the lens of justice and intersectionality, and she hopes to get into the global policy field of AI and technology and is currently doing research on democratisation of data and data feminism in regards to the global south. We’re so glad to have you with us, Sakthi.
Sakthi
Hi, Vandita. Hi, Sanchi. It’s such an honour for me to be here. Thank you for having me.
Vandita
The pleasure and the honour is all ours. We’re so glad you’re here with us today. So, I’m gonna jump right in with my first question, right? Sakthi, for me, I’d love to understand how you got involved with your work in data feminism.
Sakthi
Well, I think that’s a very important question, and it’s something that I often ask myself, as well as to what the starting point is, but then I think I can never really pinpoint the day that I start looking at data feminism. I think from when I was in school there was always a notion of justice and ethics that was always underlying a lot of what I did. I graduated out of engineering, like I did computer science and engineering as a bachelor’s degree. To be honest, initially, I did not want to do engineering. I was sort of leaning towards doing law, but then I don’t know if it was fate, but I ended up, you know, doing engineering and I graduated last year. And even while I was in college, whatever projects we did or be it the programs, the way that, you know, we were taught about it, algorithms, modules on ML, AI, big data, cloud computing. It was very objective and you know, in the usual sense they were good lessons. We got a lot of details on how these technologies work in the world in general, but one thing that always irked me is how these technologies are going to be translated for impact that’s worthy, right? Impact that’s inclusive in nature and that’s a question that I never got an answer to.
And after I graduated from our college, I think during the pandemic, it really gave me the headspace to further interrogate, read books, talk to people and that’s when I came across data feminism, the book that you talked about in the beginning, Vandita. And I also came across a lot of wonderful articles online, a lot of great books actually, and that sort of just, you know, really got me into thinking more about this whole concept of data feminism. And professionally, I’m a machine learning engineer, so it’s very much relevant to my field if I should say so. I work with data on a day-to-day basis. And again, even, you know, be it my personal project, be it professional, be it anywhere, when I talk to my peers or colleagues, the machine learning field or the data science field is looking at data in a very objective way, but I think slowly people are coming to realise that data cannot be completely objective and there is subjectivity to it and it’s very important to look at ethics, so on and so forth. I think this is very obvious from the whole talk of the US interrogating all the big tech companies, with India, trying to get in the PDP bill and so on. So, there is some form of traction happening now but I would say, in summary, in my mind, the threat of justice, the threat of ethics was always there and having joined professionally the field of data, it just made sense for me to be involved in it and really, you know, live through the beliefs that I have around intersectionality, ethics and justice.
Sanchi
Thanks so much for that, Sakthi. Thanks for sharing your experience with us and it’s so great to see that your experience was also informed by subjectivity, which is also one of the founding principles of black feminist theory, which also informs a lot of data feminism. So thanks for sharing that with us. And you know, we were talking about this briefly before as well, and we would love to hear from you. Why do you think data feminism is something that everyone should be talking about? Why is it so important, especially today?
Sakthi
So, data is not neutral and that’s a belief that I strongly hold and with advancing technologies both in terms of software and hardware, our world is increasingly leveraging data as a sort of ‘fuel’ to run an abundance of business activities or day-to-day tasks and so on. So, it’s pretty much present in everything that we do on a day-to-day basis, right? Like from our phones to the recommendation systems in the shopping carts of Amazon or like Netflix. So, some form of processing of data is continually happening and this has become sort of the advancing tech that’s been there, that started like a couple of years ago and it’s still going on, right? And I think in such a situation when the world is heavily relying on data when it’s becoming a political and a social sense, it’s very important to start paying attention to it far more than from, you know, like Vandita mentioned before, like an extractive sense. For instance, if I have to quote, like, the importance of data inclusivity in a very, you know, simple day-to-day or research-based task. Like for instance, if you take seatbelts, right? So, cars have been designed mainly on data collected from car crash dummy tests using the physique of men. As a result, women are 47% more likely to be seriously injured and 17% more likely to die than a man in a similar accident since women’s body shape and pregnant bodies are not compliant with the so-called standard measurements. So, you can see over here, how important and how, you know, data inclusivity or the type of data who we include in data research, which problem statements we choose to use data for, really matters.
And it’s just becoming more and more relevant as the days pass by when we have you know programs like facial technology being released in the public or when we have, you know, scary stuff like MobiKwik data leaks that recently happened. And also, maybe another event, I guess to quote, would be the very recent Uber driver facial recognition failure. Like, it’s very, very clear of the obvious effect of data on our everyday lives and I think it’s high time that we come out of the sort of capitalistic or like masculine nature of how people look at data research projects and to really focus more into a feminist lens, which includes collaboration, which includes a really intersectional aspect to it. So yep, I would say, I guess if you just take a look at the news or if you open social media, the answer is right there. Data feminism is something that we really need.
Vandita
Thank you for sharing that. I think that’s such an impactful point that you’re making here. Like, you know, when you’re speaking and I’m thinking about this and I’m thinking is everything I watch now on Netflix dictated by what I decided to watch two years ago, and then Netflix decided that’s all I am, and now I’ve just been watching that because those are the recommendations that come up. Like how much of me is being shaped also by the very rigid idea of what data creates of us, right? Like it doesn’t allow us the fluidity, it doesn’t allow us to evolve and change and then it pushes back on us in a way that forces us sometimes also to stay the same. Even the Uber example you brought up. That’s very interesting because it also brings to light how much of a digital divide there is and how such principles and such forms of processing data can affect different people so differently. Like in the Uber incident, a man went on a Hindu pilgrimage and he shaved his head and came back, and then the Uber new system for like taking a selfie and identifying yourself as a driver just could not recognize him anymore, but this also then speaks to the fact that because he’s someone who probably is not as comfortable with technology, he’s been physically visiting the Uber office day on day, and nothing’s happening about that. So there are so many interlinkages of issues, not only in terms of how our data is processed, but even in terms of, like, how many of us know how to get it corrected? How many of us have that sort of access? How many of us even understand these issues, right? Because I know it took me forever to be able to reach even 1% understanding and that’s where my next question even comes from. That, and it’s something that I’ve been thinking about a lot, is like, how do you think of conversations in the text space and how do you think about solutions for accommodating different types of bodies in the digital space while these differences are very apparent in our cities or physical spaces, right? Like, how does a male, white, cisgender, heterosexual body navigate a physical space as opposed to maybe a trans person or as opposed to someone else? And that makes me think about how you do this in the digital space because it’s so much trickier. So how do you think operation manifests itself in the digital space and how can we adjust it?
Sakthi
I think if you look at the way business models are even made today, they embrace the concept of rationality, self-interest and maximising efficiency. And you know and if I take that into consideration, then most of the business models cater to the majority population over a so-called minority. So, I think in the digital space oppression really manifests in the way of algorithmic bias in the sense of you know, not really having those voices or those you know characteristics really included in it purely because the business model that’s been foundationally established, the so-called capitalistic business model that’s been established, may not necessarily incentivize that, if I may say. So in that case, I would really now bring to focus the point of funding and so when it comes to data projects. So, it’s not also just about algorithmic bias, like I mentioned before. Another aspect to also take into consideration is funding and we all know that the sort of funding that is highly popular these days has the sense of power attached to funding. Like if someone were to fund me, then I know that there’s some form of power they sort of influence over the projects that I do or the decisions that I make, and I feel like this really translates into a lot of data research projects and that it becomes very important for one, to really, you know, step back and understanding what are my influences or which are the kind of influences coming into play when I’m sort of doing a particular problem statement, right? And often at times there’s also this sort of, I don’t know, addiction towards high accuracy. Like we have all these objective metrics like accuracy, precision, recall, so on and so forth.
But then sometimes I feel personally, there isn’t a metric that can really measure the actual impact of it in the real world as such, and when I say impact, like the impact on like the kind of, maybe the algorithm brings out some form of injustice and that injustice isn’t really measured, right? And there’s this very famous line that I keep saying a lot like ‘if there’s no data in it, then it never existed.’ And there’s this need for always quantifying something and I think that also comes down to the theory of burden of proof. So if there is injustice, then I need to really have a proof of it quantitatively or in some form of data somewhere but if you don’t have these mechanisms for algorithms or data systems to measure these quantitatively, then so often it just becomes invisible and no one really hears of it, right? What makes me happy though is the fact that there are a lot of organisations coming forth and talking about this, like One Future Collective themselves. And we have foundations like the Internet Freedom Foundation and a lot of authors like Sasha who wrote the book Design Justice, which is also an incredible book talking about all these oppressions that takes place in the digital space, I believe, as well as in outside spaces like in the physical space as well. So, I would say that it’s a journey, it’s a process of learning and unlearning that has to take place here. Like I really believe in education and in creating awareness. So, in colleges, I would love for engineering colleges, or even like data is also just not about technology, right? Like data is present in economics, in the sociological sector, psychological sector, so it’s present everywhere. So, it’s also not something that’s just bound in that so-called text space. So, I would really hope that we have better education or better discourses on the concept of justice and ethics in this perspective of data and all the subjects so that that is always like the underlying thread in any student’s mind or in any professional’s mind when they come forth and do work in real life.
Sanchi
Thanks so much for that, Sakthi. So many interesting things that you were talking about coming up. Like how economic capital translates into power in the digital space or you know, when you were talking about how do we quantify injustice and how if there’s no data, it never happened. It really just took me back to Mimi Onuoha’s piece on the missing data sets and how operation is so deep-rooted that what even deserves to be quantified is the first question, right? And Sakthi, since we’re talking about oppression in the digital space, and I know that you have done some work with this, why do you think it’s important to look at data feminism through the lens of the global south as well?
Sakthi
That is such an important question. So, the production and dissemination of knowledge is greatly shaped by power relations, right? And that’s where I talk about how I really believe education’s important and if you take a look at our textbooks and our curriculum, I believe that Western theories and works feature more prominently in our syllabus, in policy-making and the general wave that a lot of people think about in our world and sometimes these voices are taken as a universal default. And of course, this is not a natural process rather, it was something that was constructed over centuries through socio-political forces like colonisation being one of them. So, in such a situation, I believe that the voices of the global south here really matter. So, it’s almost like taking a stand against what we consider the norm and really relooking at everything in perspective or in the perspective of who we are and where we really are centred in, right? And there’s a brilliant paper by Payal Arora called The Bottom of the Data Pyramid, Big Data and Global South and in the paper, she basically talks of how there’s this sort of scepticism and caution on the social impact of big data in the West, you know, for instance, we had like GDPR come in, which really took into consideration the privacy of personal data. But then if you look at the global south, there’s always this bias in framing big data as an instrument of empowerment. So, there’s always, almost this sort of dichotomy that exists, and it is true that you know, data can help us in a lot of ways when implemented correctly but in the paper, she talks about analysing these developments through the lengths of database democracies, identities and geographies to really, you know, ask the critical questions of these projects. So that’s the paper definitely that everyone should read because it brings out a lot of interesting points and also like, you know if you look at big tech companies like Facebook and Instagram and Google, so on and so forth, so these are social platforms that a lot of us use on a day-to-day basis, but a lot of their policies are sort of centred around global north-based policies, right? And then that’s when it’s very important for us to really come forth and talk about the relevance of the context here, the relevance of the subjectivity there because something that may apply there might not necessarily apply here. And, you know, it’s really important to take into consideration the voice of the global south. I mean, again, data feminism is about this whole aspect of intersectionality and getting as many voices in, the voices who are truly part of the process to come in and I think that’s why it’s very important to look at data feminism through the lines of the global south.
Vandita
That’s super important, Sakthi. I mean this, you know, that takes me back to teaching about what is it that I have lost in my understanding even as I try to learn because that just maybe is not enough knowledge production or enough focus even on writers, researchers in the global south or even on practices from here, which is also to go back to the term and think, you know, who defined global south and for us to question that and that’s something perhaps we can take up in another episode. But since we are winding up, I have a question for you which is for our listeners, what are maybe like, you know your top three, top five, how many ever, active practices around data feminism that our listeners can incorporate into their everyday lives?
Sakthi
Thank you for that question, Vandita. I would say that these seven principles of data feminism are super important and they really inspired me in my day-to-day lives and the tasks that I do. For instance, if I have to take the first principle being examining power, I’m making it a habit to really sort of question and think a bit mindful of a lot of actions that I do or the actions that I observe out in the real world and one of the tools that I use to sort of think that way is by thinking of the matrix of domination.
If I have to explain the matrix of domination a bit more, it basically tries to look at a particular thing through 4 matrices. The first being the structural domain, second being the disciplinary domain, third being the hegemonic domain and the 4th being the interpersonal domain. And this comes off the belief that people experience and resist oppression on 3 levels: the level of personal biography, the group or community level of the cultural context created by race, caste, religion, gender, and then of course on top of that we have a systemic level of social and political institutions. So, I really try to look at these three layers, you know, understand what is the systemic oppression that has existed here, what is the community domain level of oppression and then what is some sort of bias or oppression that comes from my own interpersonal domain so-called.
And then one of my favourite principles of all here, in the seven by Catherine and Lauren is make labour visible, right? So that is something that is so important like you know giving credit where it’s due and also making sure that it’s recognized and valued. In the global south, I think it really matters because a lot of times you have like the so-called invisible labour. So, I always see labour in terms of this big iceberg where you only see like a peak of it outside but you have so much, you know, hard effort being done by millions of people that often goes discredited. So making labour visible is something that I do in my day-to-day life. So, whenever I do a project like even you know if someone remotely helps me and contributes then I really make sure to credit them wherever possible. So, I think the principles themselves provide a very good foundation for everyone to think of and get inspired by and do their own take on it in terms of their day-to-day activities. As long as you know at the very end, you just keep intersectionality in your mind and you make sure that you’re really inclusive in what you do, I think those actions then really matter at the end.
Sanchi
Thanks so much for sharing that with us, Sakthi. I think the matrix of domination, like everything else that Patricia Hill Collins has written and spoken about, is really just a superb way of starting to question and then challenge things around us, right? And thanks for speaking about how we can make labour visible. The small actions like you said, if somebody has helped out in a project, making sure that we recognize that and how it’s just so feminist at its core, right? I have learnt so much from this and I think our listeners would definitely agree to that.
Sanchi
Thank you so much for joining us today, Sakthi. It was so great talking to you and learning with you.
Vandita
I just wanna add that Sakthi, it is always such a great learning experience listening to you talk about this, and it makes me realise every time just how much of an everyday part of my life data feminism should be because data is a part of every minute of my life and I think we bring that to life so beautifully. So I hope for all of our listeners we go back recognizing that every minute that we exist now, we’re somewhere helping someone collect data and it is upon us to also become more mindful, ask questions and hold people accountable when it comes to data collection, processing and the usages of data. Thank you for joining us so much today. As always, I learnt a lot.
Sakthi
Thank you, Vandita. Thank you, Sanchi. It was such a great experience. It’s actually my first podcast. So I think this is something that I’ll, like, keep in my memory book and like, be super proud of. So thank you for giving me this opportunity.
Vandita
Until next time, take care of yourself and stay with us in our journey towards a radically kinder world.
***
For more delightful discussions on practicing feminism and fostering communities of care, check out the other episodes of the Nurturing Radical Kindness podcast! Until then, here’s a reflection activity for you to mull over.
| Reflection Activity – Data Feminism Take a moment to reflect on a specific task or situation in your life and deconstruct it using the four matrices discussed in the episode. First, consider the structural domain—what role did social and political institutions play? Next, explore the disciplinary domain—how did community or cultural norms influence it? Then, think about the hegemonic domain—how might dominant narratives shape the situation? Finally, reflect on the interpersonal domain—how do your own biases or personal experiences factor into the outcome? |
About the Nurturing Radical Kindness Podcast
Radical Kindness is the ethos and practice that forms and informs One Future Collective. It guides our constitution as an organisation and is the core value that guides our work. It is a politics of love, fighting against apathy and hopelessness. Often being ‘hard’, ‘stoic’ or ‘rigid’, is considered crucial for social change, and it is this very notion that radical kindness challenges. It espouses that being kind, compassionate and loving in our activism can still pave the way for dissent, defiance, growth and rebuilding. It is a tool we seek to use to rebuild our systems with care, nurturance and justice at their core. It allows us to hold various stakeholders, including ourselves, accountable in how we interact with ourselves and our communities and to build towards a lived reality of social justice collaboratively.
Hosted by Sanchi Mehra and Vandita Morarka of One Future Collective, this podcast attempts to unpack what it means to be radically kind and how we can practice it through conversations with members of the One Future Collective community.