Data Futures, AI Futures: The Scotsman Data Conference 2023

"Data Is Not Neutral"

The first keynote speaker, researcher, broadcaster and author Stephanie Hare, gave a thought-provoking talk entitled "Data Is Not Neutral". Dr Hare covered a number of ethical issues associated with data, ranging from the environmental and resource consumption impact of data centres to the need for regulation of AI technologies to bring them in line with many more tangible technologies (drugs, for example). 

Hare cited that a typical ChatGPT interaction is estimated to require 500ml of clean water, and called for all data centre operators to be open about water and electricity consumption of their facilities. Hare also raised the potentially disturbing idea of whether data collected on a national scale would be safe were there to be a regime change. Could the good intentions of governments committed to preserving privacy be undone in the future by those with a different political view?

Data and public services

EPCC’s Director, Mark Parsons was one of the experts for the panel session "Future public services: improved and driven by data", joining Layla Robinson of Research Data Scotland, Manira Ahmad from Public Health Scotland and Dr Ken Sutherland from Canon Medical Research Europe Ltd. 

The discussion focussed on recent progress in Scotland to get more value from data derived from public services, and the improvements that can be made. Mark Parsons spoke of the benefits of using the data "safe haven" model, which brings data outside the NHS firewall and makes it accessible to the research community, allowing approved studies to be done in a secure computing environment.  

Layla Robinson described the work of the newly established Research Data Scotland (RDS) to support analysts in the complex pathway to gaining approval for studies and access to data. The ultimate aim of RDS is to advance health and social wellbeing in Scotland by simplifying access to public sector data. Manira Ahmad identified the need for a lighter touch means of providing data to researchers to gain more value from public data assets and enable questions to be answered more quickly, while always preserving privacy. 

The need for improvement here was illustrated by Mark Parsons, citing an instance in which a research team needed to complete 937 permission forms in order to proceed. The panellists agreed that raising awareness of the privacy preservation strategies already in place, together with examples of the benefits of studies (for example understanding how climate change affects health) will be an essential part of the way forward.

The panel session was of particular interest to me, working as I do as part of the ADR Scotland team at EPCC. I was encouraged to hear the view expressed by Manira Ahmad that we need to simplify processes if we are to gain significant benefits from the valuable data assets that Scotland holds.  

My experience is that the highly risk-averse environment that currently exists hampers opportunities to automate processes, thereby slowing down access to data by researchers. The solution to this, in my view, is to undertake an analysis of the measures that are in place to preserve privacy, not only in terms of how effective they are in reducing risks, but also what they cost in terms of added complexity and ease of implementation. I strongly suspect that some of the measures give very little added protection and incur a heavy cost to implement. Raising awareness of the balance between risk and reward within the population as a whole is, as the panel recognised, a vital part of this whole process.

Further information

Article by Layla Robinson, Chief Partnership & Strategy Officer at Research Data Scotland: "The future of public services: improved and driven by data".

Related articles published in The Scotsman's Data Conference supplement.

The conference was sponsored by The University of Edinburgh, the Data-Driven Innovation initiative, Research Data Scotland, and Smart Data Foundry.