Why "ethics" of digital data?

This is the first of three posts leading up to the Ethics of Data Conference at Stanford. I've posted some earlier thoughts here.

In early August, the New York Times ran a "Room for Debate" series about big data. Over the course of a week columns noting the dangers of digital data use ran next to columns extolling the utility of digital data for community improvement. Other columns in the series, which was titled "Is Big Data Spreading Inequality," looked at how data may be used to extend credit to underserved communites, while others noted that data may be limiting job opportunities. One columnist called for due process related to digital data - an idea that has been studied and proposed by other legal scholars.

Clearly, there's a lot of good and bad that can be done with digital data. How we use it, what bounds we put on its use, what rights we protect regarding its use - these are classic ethical questions now being brought to the forefront where large sets of digital data are concerned.

On September 15 the Federal Trade Commission is hosting a conference in DC on "Big Data: A Tool for Inclusion or Exclusion?" which will focus on how consumers are affected by the uses of big data.

On September 15 and 16, Stanford Center on Philanthropy and Civil Society's Digital Civil Society Lab is hosting a conference on the Ethics of Data in Civil Society. We're focused on what the rise of digital data as a resource and digital infrastructure means for private actions with a public benefit.

The questions addressed in the New York Times series and those that the FTC is asking about discrimination certainly matter for civil society. Two key conceptual resources for the Stanford conference can be found here:

• Framing Document for the Conference
• Provocations on Civil Society Practices and the Ethics of Data

We'll be looking at:

• How data are being used to frame the issues on which nonprofits and voluntary associations work and what civil society can do about it;
• The realities of association and expression in a digital age and what these changes mean for civil society
• How scholarship is changing in a digital environment; 
• The rights of those being served by nonprofits and civil society;
• Ethical dilemmas for civil society organizations using digital data and how to work through them; 
• Ethical ways civil society and industry sources of digital data can work together 

Participants include activists, data companies, nonprofits, scholars, and funders. Conference resources are available here and we'll put post-conference materials there as well.

Cases for Discussion - data, ethics, and civil society

Last year when I wrote the Blueprint 2014 I included a few scenarios regarding nonprofits, foundations and the uses of digital data (see pages 22-23).

Now we're on the brink of hosting an entire conference on the Ethics of Data in Civil Society.

We're pulling together resources from the field and from scholars. I was thrilled when Jeff Raderstrong, now of Living Cities, sent me these Case Studies he and a colleague developed for classes at George Washington University.

   Cases for Discussion BigData by Lucy Bernholz

We'll share these on the conference website (http://stanford.io/ethicsofdata) where you can also find a provocation piece written by Andrew Woods -  Keep checking there for more resources as they become available.

You can follow the conference planning and pre-conference discussions on Twitter at #EoD14.

Thanks for sharing, Jeff and Katlyn.

A new calculus for civil society

In the analog era, when you took action for social good the math was relatively straightforward;
                                 1 action = some result.

It might have been a 1:1 relationship: 1 donation = 1 meal.

It might have been more than that: 1 volunteer act = multiple additional donations = multiple benefits

It might have included a multiplier effect: Action sustained over time = new policies = multiple benefits

It might have backfired: 1 action = negative result

What it didn't have, and part of what makes digital action different, is a "digital differential."

Here's what I mean. In the digital environment, every action creates a digital trail - data, metadata or both. So if you click to like the ALS Ice Bucket challenge (for example, to be au courant) your click supports ALS research.  It also tells your friends you care about ALS (or that you are very much in the know), it also adds to the dataset of information about you that is being built by several enterprises and held in several places, and it also adds to the dataset of the ALS campaign.

Your single click becomes a digital data point with lots of potential other uses (marketing, donation solicitation, friendship building). One action = lots of derivative uses and interpretations, some by you, most by others.

This digital differential is true for all digital data. Our actions in digital space create an additional "resource" (data) that can be used in lots of ways. These digital differentials may be used for positive or negative actions. What happens with them is not inherent in the data, it will depend on what we do with the data and how we do it.

CrisisTextLine is a great example of this. It helps crisis counseling centers reach teens via text. In the analog age (last year) when this was done by phone, the math was straightforward:

ANALOG: 1 call = 1 teen helped.

Today, the math is different:

DIGITAL: 1 text message = 1 teen helped + a dataset of digital text messages (with more than 3 million records to-date).

This is one way (there are others) that digital changes the calculus of civil society.

What do you do with that dataset? How do you protect it and the rights of the people represented within it? CrisisTextLine hopes to make it useful to scholars and policymakers. You can see their work - and their ethical decisionmaking, struggles, and open questions about this here.* The upcoming Ethics of Data conference will look at these questions and many others in a broader civil society context. Some of the thinking on data philanthropy also addresses these issues.



*At the Digital Civil Society Lab at Stanford, we held a charrette around CTL's work - because their opportunities and challenges are all of ours, they're willing to share them publicly and ask for help, and we all stand to learn a great deal from what they are trying to do and how they are trying to do it.

The last 20 years and the next 20 years

I had a great conversation this morning with Sara Davis at the Hewlett Foundation. It led to these two napkin sketches of the past and the future of organizational structures - what do you think?*

The last 20 years in nonprofits and foundations:

The next 20 years in nonprofits and foundations

*The drawings are mine and should not implicate the Hewlett Foundation in any way.

Opting in in digital civil society

One of the defining features of civil society, the independent sector, the nonprofit and philanthropic sector or whatever you want to call it...is that participation is voluntary. Giving your time or money is your own choice. Partaking of services offered, for the most part, is by choice (less so where government services are contracted through nonprofits, but stay with me here). One of the most common names for nonprofits is the "voluntary" sector. We are neither required by market forces nor obligated as citizens to participate. This notion of "voluntary" access is complicated when we are talking about people with few choices, those marginalized by prejudice or finances (or both). Those nonprofit organizations that serve these communities are often actively concerned with respecting the individuality and power of the people they serve - a concern that needs to be attended to every day in every action, small and large.

Voluntary participation is what allows civil society to be the home of expressive acts, protests, and the rights of minority groups in a majority-run democracy. A space, such as civil society, that is defined by our individual ability to participate voluntarily is also defined by our ability to choose not to participate. I don't have to give to causes I don't support or to organizations whose agendas I disagree with. In the language of today's tech world, the space in our democracy that allows me to "opt in" also provides me a space from which I can "opt out."

Civil society needs to take these qualities of individual power and "opting in" seriously into the digital age. Our work with digital data needs to align with our organizational values and missions.

If civil society claims a role in pursuing social justice than it has a special obligation to do two things - protect people's power to act and make sure that digital data aren't used to exacerbate existing power differentials.

Most of what is offered online, certainly by commercial operators, is "opt out." By default, your information is gathered, stored, shared, mined, spliced, diced, and minced. Anyone who has tried to cancel a Facebook account, build an app that doesn't automatically collect location information, or even make a donation without giving their email address knows that the default choices in the digital world are not "opt in." Rather, the most common default online seems to be "make it as difficult as possible to opt out."

We risk losing a great deal in a society if we lose the power to "opt in" and its "opt out" corollary. It's not just about tools like social media or text messaging - more and more policy decisions, corporate pricing choices, and infrastructural access decisions are being based on decisions made from digital data collected via remote sensors, satellite imagery, and massive datasets collected from data we leave behind us in our daily digital activities. These datasets drive decisions on pricing and access to services such as transportation, housing, child care, clinic access and so on - and they're being built with our data and without our consent (or even awareness). 

People need the power to choose. Precisely because a nonprofit may be the only service provider in an area, or the only culturally competent association, or the only option of any kind - it needs to be attentive to the power arrangements it puts in place, seeks to mitigate, or actively fights against.

How data are collected, analyzed, stored, shared, secured, or destroyed are acts of power. Commercial firms (for the most part) have made it clear that when it comes to digital data, they hold the power. I'm willing to bet you didn't read the Terms of Service for that app or website before you logged in, but that's what they say.

Civil society needs to attend to the expressions of power embedded in digital data. It's not just how the sector uses digital data, but how digital data is used to define the sector.

Ethics and the Age of Digital Assumption

It turns out that Facebook is manipulating your news feed and OK Cupid is playing with your heart.


Neither of these digital algorithmic head games came as a surprise to folks in the digital know, but that doesn't make them OK.

And they did come as a surprise to most of us -  who don't quite understand how software works and who still make assumptions about personal agency, control, and validity based on analog experiences. This is a world of "black box algorithms" - where software code is determining what information we see and don't see. It's a world where consent means "not really" - in which, as one wag puts it, "the biggest lie on the internet are the two words "I Agree." There's a new book out, Virtual Unreality, that makes it clear just how prevalent this "manipulation by software" really is. Be clear though - the phenomenon is not new, just the broader understanding that it is happening all around us.

At a recent innovation conference hosted by Stanford PACS and the International Red Cross and Red Crescent Society I called this moment in time "The Age of Digital Assumption." By that I meant a couple of things:

• For nonprofits - assume your volunteers will use digital tools (probably their own, as in "Bring your own device"). Assume someone, somewhere will create a new digital tool or apply existing ones to whatever crisis you are handling (see #DetroitWater). Assume you will not be in control of the way digital information is used to respond to any particular event or disaster.
• For people - assume your digital information is being collected, and that you don't really know what is being done with it. Assume that your privacy does not rank as high on the list of concerns of the organization collecting your data as it might on your own list. Assume that if you are giving information (or permission to gather information from your phone) such as your name, email, location data, etc. that the organization asking for it is, in fact, collecting it and doing something with it (maybe just storing it till later)

People have all kinds of feelings about Facebook or OK Cupid using their data.  One good outcome of these high profile corporate acts is that they've brought ethics into the broader general discussion about digital data and platforms in society.

Moving the issue of ethics from the fringe to the center is also part of the "Age of Digital Assumptions." This age comes after what I might call the "Age of Digital Exceptionalism" in which we were still agog at all the things our "shiny objects" can do. Now we get to the harder questions, namely, "Just because we can, should we?"

We're putting ethics at the center of the conversation in an upcoming conference hosted by Stanford PACS Digital Civil Society Lab, the Harvard Humanitarian Initiative, and the Brown Institute at the Columbia School of Journalism along with DataKind, PopTech! the Responsible Data Forum and a growing list of other partners.

Here's the basic question - What are the ethical parameters for civil society organizations using digital information and infrastructure?

Nonprofits often talk about trust and integrity as being central to their work. If so, do we want (expect) nonprofits to act differently with our data then we expect Facebook to act? Do we want (expect) nonprofits to communicate more clearly to us about what they do with our data? Do we want (expect) to be able to access civil society organizations and their services without compromising our own data (or the digital whereabouts of all the folks whose addresses and phone numbers are stored on our phone address book?)

Similarly, do we want (expect) nonprofits to access and use remote digital data in different ways than other enterprises? If wealth managers are using satellite imagery for portfolio management, does that mean nonprofits can or should use access to the same data to determine levels of poverty? Or to find ways to better deliver medicines to remote areas with bad roads? How should funders think about ethical issues, and what role do they play in shaping behavior?

What, if any, incentives should be provided to encourage civil society's use of digital data for good? What, if any, limits should be put on civil society uses of digital data?

More about the conference is online here. It's by invitation only because of space limitations, but we'll be sharing resources on the web site and conversation on twitter via #EoD14.