Thread by @steveparks

I’m at the @ODIHQ members evening, on How To Be Trustworthy With Data.
Very relevant to some of our current projects - but also to our sector as a whole.
I’ll keep some notes in this thread...
#ODIMembers

First up is @RooMarcus giving an overview of the ODIs thinking in this area.
They recently did some research with @theRSAorg which found public demand for an ethical approach to their data was strong...

They found that the public want greater...
Honesty and Transparency
Agency and Control
Compliance and enforceability
Rights and responsibility
Context and fairness
...over how their data are used

The ODI has therefore come to this definition of what data ethics is...

...and they’ve developed the Data Ethics canvas, designed to be used by all roles involved in projects involving data, from product owners to data scientists, to help them think about the issues involved
theodi.org/article/data-e…

Next up is Dominic from @colouringlondon who have used the Data Ethics Canvas in making design decisions about their service...

They’ve even published an analysis of how they meet each factor in the data ethics canvas: pages.colouring.london/data-ethics

This led to deciding not use 3rd party authentication services, eg social media, and even to make entering an email address optional (with caveat that if you forget your password you can never get back in!).
Interesting to see the canvas being used in design, not just an audit

Next speaker is David Wright a consultant on data protection and ethics.
He says the risks are growing - AI, facial recognition, and more.

But also the demand for ethics is growing. Led in part by the EU, but also by employee activism, consumer activism and more.
Under all that though - it is simply ‘the right thing to do’

He says that clearly having good data ethics is now a key factor in being able to more easily recruit the best data people.
But there also some sticks as well as carrots - the European Commission is actively regulating, there’s media scrutiny and more

He talks about ‘ethics by design’, starting with principles, and considering these during the whole agile development process. The tools can be ethics impact assessments, ethics touchpoint tables, codes of practices, and an ethics board

He says that there need to be a range of things you check, but it can’t just be a checklist that a computer could complete for example, it needs to be an exercise in using them as a prompt to think around the issues, discuss and address them

Next speaker is @robmckendrick head of data science engineering at the Co-op.
He says they focus a lot on how to be trusted with data

Some examples of the projects they looked at with regard to data ethics:
The app to help employees track their shifts.
A digital service to provide all the data for funerals in one system.
A service to provide personalised coupons to members.
Moving all websites to one platform

Part of their approach has been to turn the discussion internally away from the governance of data, to the ethics.
All projects now have to incorporate this ethics approach, based on a playbook they’ve developed and the ethics canvas.

He says there is a challenge in finding right balance of how open and transparent to be about ethical use of data - because there’s risk of press misunderstanding what is actually good practice if they stick their head above parapet and say they have issues they are working on

...and that rounds up the speeches at the event. Thanks for following along.
If you want to find out more about the Data Ethics Canvas it’s theodi.org/article/data-e…