22 Questions for ethics in data and AI
An exploratory framework to aid the design of more ethical data and A.I. products.
Ethics is all about asking the right questions in order to navigate complex scenarios with no obvious right or wrong. As we head into a data driven future, where algorithms will make what could be life or death decisions, the need for solid ethical foundations comes into sharp focus.
These questions are a form of framework for approaching ethical product design, but their underlying intention has been to act more as a meta framework — to assist the creation of more specific ethics policies.
The Structure
The questions are divided into three levels. The first level explores the creative process, the second level looks at questions to ask while actually implementing and managing a product, while the last level looks at the organization and governance.
Question zero is essentially a framing of a, perhaps, lofty goal: To foster an organization that will, by it’s nature, make strongly ethical choices.
22 Questions on data ethics (V1.1)
Big picture
0. If anything was possible, and there were no constraints, how would we act?
Design stage
- What is our intention?
2. What are the components we need to bring our project into reality?
3. How will we learn about the potential consequences of our work?
4. How will we establish, and maintain, authority and trust?
5. How are we expected to conform?
6. Who are the participants in our creative process?
7. How will we cede control to, and support, those responsible for day to day choices?
Implementation/Management
8. What is the balance of power between those represented in our data, and those possessing it?
9. What does our data potentially reveal about the personal lives of those represented in it?
10. What events, beyond our control, should we be planning for?
11. How will we be aware of, and adjust for, bias?
12. What, or who, are we willing to sacrifice, at what price?
13. When does data get deleted?
14. Who, or what, else will be involved, and how do they align with our values?
System/Organization
15. What forms of behaviour do we actually incentivize?
16. When we fail to live up to our standards, how will we respond?
17. How will we provide guidance on how to navigate difficult issues?
18. What aspects of our work would we like to keep hidden?
19. How much of our inner workings do we reveal to the world?
20. How will we encourage, listen to and incorporate feedback?
21. How will our ethical policies continue to grow and evolve?
Although these questions have been brewing for a while, I got the chance to get them fleshed out developing a workshop for a major multinational firm The principles that they adopted for as their ethical foundations all emerged from the workshop, so I have some confidence that we are on the right path — but I have also come to realise just how much I have to learn.
Together with our partner, Signify we have developed a two day workshop for organizations seeking to develop ethical products with data and AI. This workshop is our way to continue to learn, and to share some of what we have learned so far. Get in touch if you want to join in.
