Michael Veale

Researcher into machine learning in the public sector.

I'm based at the Department of Science, Technology, Engineering and Public Policy (UCL STEaPP) at University College London, where I am a doctoral researcher focussing on understanding what responsible machine learning in the public sector might look like. Machine learning algorithms learn often complex and subtle patterns from data that may elude or overwhelm other methods. They can use this understanding to answer queries about the areas 'learnt' about, often — but not always — displaying remarkable accuracy. Machine learning is being used across the public sector, in areas from tax to policing to justice. The flexibility of these tools is a double edged sword. While it means it has the potential to be a widely transformative technology, helping with a broad range of problems, it also means that there is no one-size-fits-all approach to responsibiity, encompassing areas such as fairness, accountability, robustness, efficiency and public dialogue.

I'm supervised by Dr. Jason Blackstock and Prof. Anthony Finkelstein.

Alongside doctoral research, I'm also an external data governance researcher in the science–policy unit of the Royal Society.

I've previously worked in the European Commission on the nexus of health, technology and ageing; at Bonsucro, an innovative metric sustainability certification for sugarcane designing data processing and visualisation strategies; and in a range of consulting roles on technology, sustainability and policy.

I tweet @mikarv, and can also be found on LinkedIn. You can also pop me an email at m.veale@ucl.ac.uk, or drop me a line at +44 (0) 20 3108 9736. A PGP public key for me can be found on keybase.io.

Recent publications

We consider the opportunities of an emerging range of sustainability standards measuring performance on dimensions of sustainability rather than prescribing particular technologies. Veale, M and Seixas, R (2015). Moving to metrics: Opportunities and challenges of performance based sustainability standards. S.A.P.I.EN.S. 8(1). https://sapiens.revues.org/pdf/1713

Selected blogs and reviews

"We must invent better ways of measuring and understanding the algorithms’ impact over time, and their interaction with the social, technical and environmental systems we connect them to."

Nesta (2016), To Grasp Algorithms in Society, Computer Scientists and Social Scientists Must Team Up

"Given even their programmers cannot rigorously ‘interpret’ the decision processes used by their creations, how can algorithms be put up for public debate?"