Douglas Namale of Map Kibera Trust shares a map of Kibera schools with a school leader.

Those who are currently on the margins of a society are in a particularly disadvantaged position with regard to decision making, and data and maps may either contribute to that imbalance or help to rectify it. Vulnerability essentially means that variations in the status quo – the fluctuations of economics, health, environment, natural disaster, war, etc. –  will affect certain communities and individuals disproportionately. My motivation for developing ethical guidelines for mapping with vulnerable populations is simply that this is where I’ve seen the most critical ethical challenges arise. If we cannot create an ethics that protects and ideally empowers the most critically vulnerable populations, then we cannot claim to have an ethical approach to our mapping at all.

As a starting point, I am interviewing a cohort of mappers, geographers, and other experts to weigh in with their thoughts. I’m also reviewing relevant details from other frameworks for ethics such as the AI and ethics literature, responsible data frameworks, and the Principles for Digital Development.

An important point: I will be developing these ethics from the perspective of what matters to vulnerable communities as much as possible. This means emphasizing things which local mappers and residents indicate are most critical for themselves and their communities. These may not be the same ethical priorities as what geographers from more developed countries come up with. 

Here are a few of the emerging themes:

1. Consent: Did anyone ask? No, usually not. Is it really appropriate to “map people”  who live in a precarious situation, and have very little power to begin with, without talking to them? If yes, under what circumstances? What might be unintended consequences? There is an interesting discussion to be had around what constitutes consent when discussing concepts which may be completely foreign to people (new technologies, data privacy concepts, and so forth).
2. Community Involvement: Beyond just asking permission: have people residing in the area been involved, perhaps via community mapping? In many cases, mapping is a data extraction process and people are neither given back their data/maps or shown the results (or much less often given a seat at the table making decisions with that map). We need to think a lot more critically about how geo-data is being used and whether it’s actually benefiting people in vulnerable situations. Do people have a right to access their map? In a humanitarian emergency, at what point do such rights begin to give way to imperatives of saving lives? Even raising these questions is not often currently happening. 
3. Opportunity, Capacity Building and Sustainability: Failure to create equal opportunity locally: If you talk to most of the local mappers I know, they’ll find this to be the most critical problem for them personally. Mapping is a huge industry at this point, but it’s not creating job opportunities proportional to its wealth in local contexts. There’s a problem when an industry is huge and profitable, but the benefits aren’t getting to the people. Can we agree in principle that vulnerable people need to be benefiting, not just companies? How do we protect people from being used as data collectors (or “human sensors”) without any type of benefit to them? In this case, there are many examples of transferring skills, but most organizations do not have a commitment to capacity development and inclusion in geospatial work, even when they have other kinds of programs devoted to this. It may be a problem with considering “data” and “maps” and “GIS” as something outside actual programming, not subject to the same mission as the rest. But, increasingly, we’re seeing that if you do not integrate data collection and mapping into your long-term strategy, you’ll fail at sustainability more broadly. Should there not always be a requirement to train someone locally in your tech? 
4. Privacy, Imagery, Drones: Close up imagery can be a major invasion of privacy. People’s lives are lived outdoors in most of the poorest communities, not inside massive houses, so drones are potentially more invasive than with other locations. What is the best practice around balancing a need for good and quick maps, and the potential for exploitation inherent in these situations? What about emerging technologies like using geospatial AI in disaster? Or anything new that will and may arise in the future? Privacy also includes responsible data practices: how do we need to incorporate these into mapping best practice? 
5. Coordination and Duplication: There is often no coordination among mapping efforts for development, creating duplication. Shouldn’t there be an agreement in principle to work together, at least among donors and NGOs? I’ve seen direct harm happen simply by wasting resources, wasting people’s time, and also ultimately creating disjointed and conflicting data systems in developing countries where such systems are imposed by external donors, each of which sees itself as trying to help. Without coordination, when we are talking about geospatial data, disparate and often competing systems are not helpful and can cause confusion and harm. 
6. Unnecessarily Closed Data: Proprietary maps are still being made, and for what? Why? To whose benefit? Who holds the keys to those maps? I am, of course, an OpenStreetMap proponent, and I often see a failure to consider the ethical implications of closing off data that should be openly accessible, public information. This is not to say all maps must be open, but those which clearly should be open should not be arbitrarily closed off. 
7. Do No Harm: This is a principle that always undergirds development and humanitarian work, but that does not mean we are accomplishing it. We might need to specifically show cases of harm and help practitioners to understand how this maxim applies in the geospatial data for development arena.

These topics are often contentious, and I’ll first look for some consensus areas. But, the interests involved are also often at odds. My intention is to bring out some of these debates and develop a living document which will help guide those looking to include geospatial data in their work, and curtail some of the common problems that we already see occurring.