#EOD14

Infusing Ethics into Data Projects

Education, assistance and enforcement are needed to build better, ethically balanced data-driven projects. The Ethics of Data workshop on the Data lifecycle discussed ethical scenarios and key aspects of a data-driven project. This lead to many sticky notes and attempts to create the big asks and some outputs. The participants came from many different disciplines, which helped us quality check and inform our review.

Our group together created 3 asks for further review:

Resources: This should be a center online for people and NGOs to share and find guides on building ethical data driven projects.
Tool/Template: There should be a data risk/benefits/costs template as part of every grant application to build better ethical projects from front to back.
Review: Could there be a non-profit assistance group that provides free, consultation on ethical questions? How would they enforce standards? Could it be like a review board for research projects?

Milestones for an Ethical Data Project:

We created over 100 ideas and grouped them into categories as milestones. There should be checklists for each milestone including key questions for the team. While the EoD team started on the mini-ethical checklists, this really needs more iteration. We also highlighted some milestones that are often overlooked or underfunded such as: a data collection checklist, pilot, quality control, verification, documentation, secondary use and impact/monitoring & evaluation. One other observation in our conversations was proper project management skills to scope many of the ethical minefields in advance of the project pilot.

*****
Without trying to influence the room, the key project milestones resemble the toolkits we created at Ushahidi. Toolkits, including the Ushahidi one, do need to include more ethical statements/checklists to improve the success of building better projects. But, toolkits are only as useful as those who use and enforce them.

Infuse with Ethical Checklists

We gifted the Responsible Data Forum with a list of 70 questions, terms and ideas for the key milestones. If all the various milestones of a data-driven project are infused with ethical questions, checklists, and recommendations, this resource would be incredibly useful. Participants suggested that the checklists start with very generic items, but be broken out into topical domain recommendations/checklist items. This is to ensure adoption and remix in diverse fields from human rights to health to science. Some of the questions that really drove conversations included:
Data analysis: What do you do when your data analysis provides negative results (from your hypothesis)? Quality Control: Can the data be re-identified? Resourcing: Who can collect the data and why? Secondary data: What is the time horizon on the data and future use criteria?

Ethics of Data – Resource Center / Review Board:

One of the outputs of our first day of brainstorming was the need to have an Ethical Resource and Review Board. The team split off to debate the pros and cons of this idea. They even determined some of the needed services such as legal referral.

More resources:

The EofD team recreated a list of additional Ethics reading. It might keep you up all night with worry, but better to be ‘in the know’.

As well, here is a compilation on Domain-specific digital ethics, practices, and conventions. The one that really opened my mind was the Bellagio Framework: “Big Data, Communities and Ethical Resilience: A Framework for Action.” They provided criteria to consider: Governance, Place, Socio-Cultural Context, Science, and Technology.

Thank you

Thanks to our hosts at Stanford Center for Philanthropy and Civil Society (Kim Meredith, Lucy Bernholz, Rob Reich and Sam Spiewak) for making the Ethics of Data event possible. Thanks to my co-host Patrick Vinck of Harvard Humanitarian Initiative for great conversations. As always, Aspiration (Gunner and Misty) supported us with facilitation that inspired a collaborative and productive event for all. And, lastly, thank you to all the participants for being so thoughtful and inspiring teachers as we all trundle down this journey to bring better decision-making to all of our work.

(Note: It is my hope that we can infuse HOT with some of this work and trial some ideas in our work. I’ll be sharing it with my fellow Board members and the HOT Community.)

15Sep

Data Cycling: From Choices to Consent

Stanford University has convened the Ethics of Data conference this week bringing leaders from industry, humanitarian, research and civil society together to discuss and build plans for data ethics in all our work. I’m participating by co-leading a workshop on the redefining the Data Lifecycle with my colleague, Patrick Vinck of Harvard Humanitarian Initiative.

The Humanitarian OpenStreetMap Team and the wider Crisismappers community are discussing and using drones and satellites to capture imagery of land changes, displaced camps and post-disaster areas. Some of the topics we are discussing within these communities include how to include local NGOs and the government. In times of crisis, humanitarians and technologists are moving very fast. We need to have more guidance, research and best practices. At the Ethics of Data event, we will use HOT as an example data project in our conversations.

The HOT community is incredibly committed to helping humanitarians and affected communities with maps. We are frequency discussing how and what types of aerial imagery be shared. What kind of training do we provide for review and use of aerial imagery? What happens to the data after the emergency? What kind of ethical code should we provide for all Digital Humanitarians? As HOT builds Open Aerial Map and groups like the UAViators come to existence, we need to consider data education and use beyond text and include video, photos and drone/satellite imagery. In the US, Mapbox is educating use of drones by creating a map of where not to fly. This work is moving faster than the research. It is my hope from the Ethics of Data event that I will be able to convene conversations within HOT to determine our next steps. Keeping in mind that every humanitarian situation is different as are the jurisdictions in which global, local and remote contributors participate.

What about the Data Pipeline?

Data Storytelling via Infogr.am

All around the world journalists, civil society groups and governments are working on open data projects. HOT is just one of many projects aimed at using open data to affect change. Determining the new data pipeline could be informed by HOT’s experiences. In the past year, I have been at countless events where people talk about the importance of open data, the importance of the data pipeline and the impact of data storytelling. I also believe that data needs to be open, when appropriate. But, I am wary about preaching about open data without including a clear ethical compass. One of the main reasons that data is not open is that people do not trust how the data will be used. And, frankly, it is very unclear what constitutes a clean dataset. Over the past years, I have worked with Geeks without Bounds, Ushahidi and Data Science for Social Good to try and solve this challenge with checklists or tools. Every dataset begets new questions. All of our work needs to be infused with questions about data accuracy and data ethics. It is misplaced as an afterthought.

Processing the Data Pipeline

This is a Data pipeline via OKFN:

We need to create communities and tools with the data ethics checklists embedded into every aspect of projects from inception to funding to creation to education to analysis and to impact assessments. I’m truly looking forward to rethinking data project lifecycles and sharing the outputs with various communities to discover and remix together.

How would you rebuild the Data Lifecycle?

(Thanks to Nika Aleksejeva from Infogr.am for the Data Storytelling diagram.)

Tag: #EOD14