Amidst a new raft of ethical data challenges that threaten to erode public trust in charities, we highlight some of the useful lessons and resources from the first year of DataKind UK’s data ethics committee.
As new digital technology emerges, it throws up a Pandora’s box of ethical challenges and dilemmas. New regulations around data protection such as GDPR have become necessary to protect people’s rights as our lives become increasingly enmeshed with the bits and bytes that represent us in the digital world.
But regulations take years to come into effect, and technological progress is happening at an accelerated rate. The world of AI and machine learning is simply a ’wild west’ of opportunity that could be in danger of running away with itself, taking us in potentially harmful directions before we have time to fully understand the consequences.
As Charity Digital Code chair Zoe Amar argues, all charities, regardless of size, have a responsibility to be aware of the potential harm that technology can cause if used inappropriately and safeguard their service users from risks. But understanding the issues can sometimes seem complex or distant from the everyday operations of charities.
That’s why raising awareness of ethical concerns around digital has become a key focus for many charity partner organisations. Last year’s Charity Digital Skills Report put questions around the ethics of digital to charities for the first time, highlighting that the ability to assess how responsible your charity is with digital is now considered a core skill.
DataKind UK’s ethics committee - A proactive engagement with ethics
As discussed in our recent podcast provocatively titled ’Is ignoring AI the charity sector’s biggest tech mistake?’ AI and machine learning technology bring many efficiency-driving benefits for smaller charities. However, they cannot afford to plead ignorance on issues such as unchecked bias and prejudice within AI algorithms that could negatively affect the people they help and how they recruit staff and volunteers.
When it comes to ethical issues in data use, this is just the tip of the iceberg. Various considerations around privacy, transparency, fairness and consent must be factored into account before any charity embarks on any data collection and analysis.
Data-for-good organisation DataKind UK launched in 2013 as a way to connect data scientists with the UK charity sector, and a year ago it launched its first dedicated committee on data ethics, with the goal of increasing awareness of the sometimes complex ethical considerations of doing data science projects.
Christine Henry, head of DataKind UK’s ethics committee says: “Part of helping social change organisations to use data is making sure that the data and its applications (such as machine learning algorithms) are used responsibly.”
“We also want to help our community of volunteers and social change organisations think about ethics in their wider work, and find practical processes to incorporate those discussions within their teams and organisations.”
And while most small charities lack the time to launch digital ethics committees of their own, there are certainly approaches they can replicate from DataKind UK and advantages to collaborating with them.
Leading the debate
As DataKind UK says on its recent blog: “While embarking on this journey of diving deeper into ethics we realised that there is no shortage of resources to learn from. As we did our own reading, we also shared what we came across, and are continuously updating this to reflect a good set of material for someone relatively new to the topic.”
The organisation hosts a data ethics ’book club’, open to anyone, to discuss developments in data and AI ethics. The book club offers a timely opportunity to explore topics related to data ethics in-depth, through books, research papers, newspaper articles and sometimes videos.
So far it has covered topics such as face recognition, fairness, financial inclusion and gender among others, operating in London, Edinburgh and online. It’s worth keeping tabs on when 2020’s meetings will take place, or you can read about previous discussions on DataKind UK’s blog.
Charities wanting to fire up the conversation by starting their own data ethics book club can freely access the resources and materials that DataKind UK’s ethics committee used in their sessions.
Another core piece of DataKind UK’s work over the past year has been running digital ethics training for its data science volunteers, providing guidance on how to recognise potential ethical issues early on.
Working alongside its volunteers, the organisation created a set of digital ethical principles to work by which they could all agree on, using variants of existing ethical data codes. In order to get the discussion going on the real-world applications of the principles, they used a case study based approach drawing from examples across the industry, charities, government and their own past projects. The end result is five key principles that it has started rolling out to new volunteers.
While this set of principles is unique to the challenges of DataKind UK’s data science experts, DataKind UK used the process to help the Association of Medical Research Charities (AMRC) build its own ethical framework, identifying key principles from a review of 60 ethical codes, interviews, and through consultation with AMRC and their members.
This has ultimately led to nine core principles that DataKind UK’s ethics committee advises charities should consider when developing their own ethical codes that address their unique challenges and values.
Our recent podcast ‘Safeguarding the sector: designing digital ethics for your charity’ may also be a good starting point for charities to start pinpointing the ethical considerations that are most relevant to them, as we look at how charities can anticipate the difficult questions that may arise from technological change.