In the newest installment of the ATM-Lab podcast series, Chief Editor, Jenee Iyer, speaks with Lucy Bernholz, Director and Senior Research Scholar at Stanford University’s Center on Philanthropy and Civil Society.
Lucy Bernholz: Hello, it’s Lucy.
Jenee Iyer: Hi Lucy, this is Jenee Iyer from Carnegie Mellon University. How are you doing?
LB: I’m good thanks, how are you?
JI: I’m doing very well. I just want to let you know, quick, in regard to the interview, I do have the call recording going at the moment so that’s just the way the system that I’m using at the moment works.
JI: Thank you so much for agreeing to chat with me a little bit about your work and um, what a digital civil society means because I found your talk at the conference very, very interesting and I think a lot of our readers at the Arts Management and Technology Lab would really find it fascinating.
LB: Thanks. I’m looking forward to it.
JI: Thank you. So, I apologize, but I’m just going to kind of (inaudible) you had a wonderful time on your trip, but I’m going to just kind of jump in quick, because it is going to record for 20 minutes and I honestly do not know what will happen when we hit that 20-minute mark, if it will stop the call or what it will do (laughter).
LB: Oh! (inaudible)
JI: It’s going to be a surprise to me as well!
JI: The first question I want to ask is, how would you define a digital civil society?
LB: A digital civil society is all the ways that we individually come together with other people take collective action, voluntarily associate and use our private resources for public benefit in the digital age.
JI: Ok, so, I know when you spoke at the Tech Now conference, you talked about how nonprofits and organizations now no longer just have to think of the cost of their operations in terms of time and money, but also in terms of time and money and digital data and that distinction being that time and money are viable goods and excludable, but digital data is non-excludable and non-viable. So, what types of challenges do you see for nonprofit organizations when they are trying to manage this type of data?
LB: So, there’s a couple of different ways of thinking about that. First of all, the digital data that nonprofits and all of us have to worry about are everywhere. There’s the deliberate information that we collect and then store and use digitally, so you can think of that as being the, um, evaluation results, or the program services, or the kind of active program management information that we use to run any organization which the minute you are collecting it and storing and using it on a laptop or on a network computer or if you have a network printer any of that becomes a digital resource to pay attention to. But then there’s also all the kinds of digitized data that we don’t really think about as being digitized data or actually even being data that directly is, it’s first purpose is the running of the organization. So that includes all of the information on website hits, the photographs that people use in their communications, the video they may capture if they host speakers and record those meetings. So, the first, most important thing for nonprofits to recognize and think about is that they are actually surrounded by several different kinds of digital data that they generate, some of which is quite sensitive some of which is not, some of which they have a lot of control over, some of which they don’t. But the very first step is to do what we call, take an inventory of that data, to make sure that you have at least awareness, at the very least to be aware of the.. of everything from email addresses to the social security numbers that you’re storing somewhere in your system for whatever reason or the home addresses of the people you serve. And then this process of triage really is prioritizing what is most critical to the organization what is most sensitive to the people inside and outside of the organization and then from there, a process of developing ways to both minimize that information and secure it while also making sure you still have the data that you need to do your job well. In the course of that process, people often figure out that they’ve got a whole bunch of stuff. They think they’re surrounded by data, they ought to be able to do their job better. Often they’re surrounded by data, but it’s not the data you need to be able to do your job better, it’s just the data you’ve been collecting since you haven’t really been thinking about it.
JI: Right, and with the idea of data and data collection I know there’s been a lot of chatter recently both within the nonprofit world and outside of nonprofits about using technology like Blockchain to store data and I know in the nonprofit world in particular people are interested in that in terms of storing their data. On your blog Philanthropy 2173 you talk a little bit about Blockchain, speaking specifically about how it’s a little dangerous to experiment with information about people as opposed to just starting with supply chains and things. So, I was curious to hear a little bit more about your thoughts on storing that data particularly in new and emerging technologies, like Blockchain.
LB: Yeah. So, you know, I think for folks who are trying to do whatever it is that they’re trying to do without becoming technology experts, things like Blockchain start to get a lot of press and people hear oh it’s private and secure well we must need that! And that’s not their fault, that’s the…But the Blockchain is not yet, well, it’s behind the scenes in some apps, but what folks need to do a better job before they start jumping at where they’re going to lay this technological solution is really get a much stronger handle on what it is they have, what the different best uses of that information are over time. They most definitely should involve the people whose data it is in those conversations. So, there’s an incredible innovation in governance opportunity here because most of the most important data that a nonprofit is going to have is from their donors, just not from their financial donors, it’s from their data donors which happen to be the people they’re serving, who they’re probably not even asking if they can have all this data. And again, depending on what kind of organization you’re talking about the permanent unchangeable nature of that data may not be in anybody’s best interest and that’s actually what the Blockchain, what Blockchain technologies provide, is a permanent unchangeable storage solution that’s distributed over many different computers and people. And there’s going to be all kinds of instances in which whoever a nonprofit is serving, they don’t want that out there. They don’t want it in a place that they don’t have control over they don’t want it to be unchanged, they don’t want it permanently stored, in which case, oops, suddenly Blockchain is the wrong technology. So, before jumping at any, you know, shiny object being dangled in front of anyone we all need to do a better job of really sort of stepping back and saying what is the information I have, what is the digital data that we’re in charge of what do we need to use it for? In most cases where a Blockchain will be sort of sold as the sexy solution just a really good database will be a better solution.
JI: I heard an interesting quote not that long ago that was talking about how we need to treat the data of a person with the same dignity we would treat that person themselves.
LB: Yeah, absolutely.
JI: I think that’s really appropriate when you’re thinking about storing that much personal data and storing it in a way that you may not necessarily be able to delete it or erase it or be able to respect the wishes of the donor with how they want their personal information stored.
LB: Right. Absolutely. And I think particularly because so few people can do a really good job of explaining the Blockchain, we are well past the point in our technological zeitgeist, where if you can’t explain how it works, you should still default to wanting to use it. We’ve been burned so many times as individuals and as a society by these planned promises of technologists who don’t understand the human context into which a solution is being inserted or it was absolutely designed for a different reason. I mean, Blockchain technologies got so popular as a way of circumventing legal authority, so if we talked about it that way I think people would be a little bit less like, ‘oh, well that must be what I want’, because actually probably you’re not trying to circumvent the law. But it just gets confused for people in terms of all the things that are sort of, you know, some of the marketing if you will of these technologies.
JI: On that same kind of thread of tech dangers, if you will, and the idea of needing to know and understand the technology that you’re working with, I hear a lot of people talk about the idea of data and algorithms being deployed to solve public social issues. So, I very recently attended a talk, and one of the speakers was making statements about how the effects of bias in data and as being almost an inherent characteristic and not a bug of data machine learning and artificial intelligence. And the idea that kind of entrenched bias and some of the biases underlying the data are not an issue that can be sort of nonchalantly brushed aside. But particularly true I think when nonprofits who are collecting either data about their attendees or their donors and how if you don’t kind of pay attention to what data you’re collecting, from where, and for what purpose, you can end up sort of re-entrenching certain beliefs or end up targeting only certain groups. So, I’m curious to kind of hear some more of your thoughts on how public serving organizations should frame their data collection and analysis both at the data collection point and then too in sharing their results of the analysis and acknowledging what bias there may be in those results purely from the way that the data was collected.
LB: Right. So, this is also one of…this is a great question because we’ve been sort of led to, we’ve been swayed by the kind of massive data analysis capabilities of a handful of very big companies that have the resources to hire the statisticians and computer scientists who can actually do appropriate analysis of large datasets. I know very few nonprofits that have that skill set in-house. So, there are some but even there, a lot of the most cutting-edge methodologies in data science are premised on petabytes worth of data that then the nonprofit might have the analytic skills but they’re not going to actually have the data, and just as if you try to draw a meaningful statistical conclusion from a sample size of 10 and it’s going to be wrong when you need a larger sample size, the same is going to be true if you throw heavy duty data science at a data set that’s too small to begin with let alone one that was assembled for a different reason. So, for the most part, nonprofits are surrounded by digital data, they’re holding onto it in all kinds of ways on their web servers, on the back of their websites, on their own databases as they think about the information they collect from program services. But they’re collecting data that they’re then either trying to retrofit into some questions and try to answer them, which isn’t going to work well. It’s going to introduce all kinds of, well both statistical bias and human bias. Or, they’ve got a whole bunch of information, none of which is the right kind of raw material for doing the analysis that they might want to do. So, in very few instances is “collect more data and hold onto it forever” the right mind-set for a not for profit organization because they don’t have the core resources to actually use it safely and appropriately. They’d be better off collecting less data and holding onto it for a defined period of time, because that way they minimize the chances that it’s going to fall into the wrong hands. So, it’s funny in the zeitgeist of our time is more data and hold it forever. Well that, you know, if you’re Facebook, that’s your mantra, and I’ll leave it to your listeners to decide if they think Facebook is the good and the right thing with that mantra. For an afterschool program or a cultural center or an animal shelter, whatever it is, chances are that is not the right (inaudible).
JI: Right, and that makes a lot of sense. That would perhaps be one of the social behaviors eking into the workplace sort of ideas.
LB: Exactly! Exactly, you know I think I often say to people how many pictures of your kids or your dog do you have on your phone? And usually they have thousands and when I ask them if they can find the one they want the answer is usually ‘no’. And that’s the problem, right? We’re habituated into collecting as much data as we possibly can because it has become so inexpensive to do so. But it doesn’t make it actually easier to just find the photo that you’re looking for. And the big companies behind those now, the Apples, and Googles, and Facebooks, they’re throwing all kinds of algorithmic capacity at the photos to help you, to show you the ones that they think you want to see. That right there is a personal experience that almost anybody with a cell phone can now put themselves through and ask themselves if that’s an appropriate way to think about the data you’re collecting on the people you’re trying to serve. Should you be making the decision, do you have the capacity to make those kinds of algorithmic decisions? No. You don’t have the skills or the data set to build those tools. And it’s a third party. It’s someone at some other place making decisions about which photos you want to see on your phone, which I find creepy enough, but totally inappropriate if you’re talking about the use of someone’s home address information or trying to cross-reference information about the kids in your program and some other larger data set without knowing that you’re doing it in a statistically appropriate way. And it is statistics, I’ll say. We call it data science and that makes it sound super sexy and there are some unique capacities to data science that are different from statistics, but most of it is statistics.
JI: A lot of it kind of reminds me and ties back to some of these ideas of digital colonization and the ideas that when you accept the data algorithms whether they’re from an organization like Facebook or another kind of social aggregator you are taking it in whatever form they’re going to give it and you may be inviting it into your home, into your organization, but at some point you’re playing by their terms and not necessarily the terms that are yours and the terms that you want to represent your organization.
JI: So, thank you very, very much. I think we’ve made it just under the 20 minutes before we hit the “I don’t know what’s going to happen mark”. So, thank you very much. I really appreciate you chatting with me and kind of sharing some of your thoughts on what digital civil society means, what it means for nonprofits, and sharing some of that with me and with our listeners because I think a lot of them will find this very interesting as they’re kind of grappling with the societal and ethical implications of the way that they’re managing their institution’s data.
LB: Great. Thanks for calling. It’s been fun talking to you.
JI: Okay, great. Thank you so much, I really appreciate it.
LB: Yeah, Thank you.