The complexity of data ethics – Global Ethics Day 2021
Global Ethics Day 2021 was on Wednesday 20 October, held to help raise awareness of ethics and that ethics matter – whether in the daily lives of individuals, the decisions of corporations or the creation of laws and policies.
Dr Tim Lowe, who recently joined the VRU as Lead Researcher, shares his insight into the importance of ethics in data use.
Ethically, the world has become a lot more complicated. This is very effectively captured in the Netflix comedy show The Good Place. In the show four people awake after their deaths in a version of Heaven and find that all their actions throughout their life have received a point score, either negative or positive. The total number of points you receive dictates whether you go to The Good Place or The Bad Place.
On the face of it, the moral equation seems quite simple. Do something nice like buy your partner flowers and you get positive points, do something negative like attending a show by the California Funk-Rock band The Red Hot Chilli Peppers and you lose points (don’t ask me why, the show doesn’t like the band). As the show progresses, however, they complicate this massively.
Whilst two hundred years ago you might get your partner flowers by walking down to a local field and picking them by hand, these days you might order them on your phone, which might have been made in a sweat shop, which uses electricity and burns fossil fuels, is delivered in a van that might further damage the environment etc etc. The same action (buying your partner flowers) would gain you points two hundred years ago but would lose you points now.
We have to consider where we source the flowers. Is there a company closer by so they don’t have to be transported as far? Which ones uses more sustainable farming practices? Should I support local business rather than big business? There are all sorts of moral problems and conundrums that exist today that simply did not exist before.
Perhaps nowhere is this more true than in the field of data. Two hundred years ago algorithms, machine learning, and so on wouldn’t even have been a distant dream. As our technological capacities grow, so does the ability of these tools to perform more and more complicated functions. Inevitably, this presents us with a completely new set of ethical problems to consider, or at least forces us to rethink our standard ethical frames for thinking about the world.
Many of these problems have been thrown into sharp light by the Cambridge Analytica scandal, which emerged in 2018. The revelations led to greater public interest in the use of data, partially evidenced by a number of documentaries focusing on these issues on Netflix that appeared afterwards.
There are concerns around privacy, transparency, bias and, in the policing context, the potential for these tools to over-police particular groups. These concerns are, of course, legitimate but sometimes these debates tend toward a similarly simple moral equation that The Good Place begins with. Using data is bad and any attempt by any organisation to use must be bad.
The fact of the matter is that it’s much more complicated than that. Yes, of course, there are a number of legitimate concerns that need to be addressed, but there are many ethical goods that can arise from the use of data as well. These goods can range from better identification of risk factors for violence, to being able to target increased support for those who are more at risk (amongst others). By being able to identify risk factors and increase support for individuals, the social and economic harm that serious violence can do can be dramatically limited, and the prospects for the lives of individuals will only increase.
In recognition of this tension between the ethical goods and harms that data could bring, and the resulting responsibility this places on the shoulders of those embedding ethical decisions into algorithms and models, the VRU has sought to provide external assessment of the ethics of its data use. At the beginning of October, the Data Ethics Committee sat for the first time to help our data scientists think through the ethical implications of their work even further. Projects can return to this Committee multiple times to ensure a rigorous and continual assessment of the work, and to ensure that decisions are not made in a vacuum, but with the continual assessment of members of our community, subject matters experts, and ethicists.
Further to this in September 2021 I joined the VRU as a researcher. It’s my job to evaluate the interventions of the VRU to help to understand whether the ethical goods that its interventions could lead to are being produced. This builds on my own background as a researcher and lecturer in philosophy and ethics, and I am also soon to start working as a Visiting Academic at the Big Data Institute, University of Oxford to help to develop these conversations around ethics and data even further in a policing context. Ultimately, the decisions of the VRU to establish the Committee and bring me into my role with an emphasis on evaluation and ethics, shows the commitment of the VRU to embed ethics, and difficult ethical conversations, at the heart of its practice. It will, as DCI Lewis Prescott-Mayling points out in his VRU blog written earlier this year, help to ensure that ‘there is no “misuse” of data, whilst ensuring there is no “missed use” of data.’
Dr Tim Lowe is the current VRU researcher and will soon be starting as a Visiting Academic at the Big Data Institute, University of Oxford. His research and teaching background is in philosophy and ethics.
The Thames Valley VRU is leading development of how data can be used to join up multi-agency responses to serious violence. Data ethics is a vital component to this project. You can learn more about the VRU’s Data Ethics Committee here.