The Internet, Political Science and Public Policy in the Digital Era

The Internet, Political Science and Public Policy in the Digital Era

In mid-May, PaCCS Communications Coordinator Kate McNeil sat down with Helen Margetts, Professor of Society and the Internet at the University of Oxford to discuss her work on The Internet, Political Science and Public Policy: Re-examining Collective Action, Governance and Citizen-Government Interactions in the Digital Era, a program of work which was designated as part of the Global Uncertainties Project by the UK Research Councils. Professor Margetts was Director of the Oxford Internet Institute from 2011-2018. She is presently the Director of the Public Policy Programme at The Alan Turing Institute.

Kate McNeil: Thank you so much for taking the time to speak with me today. Would you mind getting started by sharing how you ended up working on the subject of the internet, political science, and public policy?

Professor Helen Margetts: The Economic and Social Research Council funded this work as a professorial fellowship – they provided me with the funding to carry out a programme of research that was not so specified in advance. The research councils can tend to be quite risk adverse, and funding my professorial fellowship was a brave thing to do, because while I’d identified various sorts of possibilities and data sources in my initial proposal, I wasn’t quite sure yet what I was going to do with them. In the end, I like to think that the programme ended up being pretty good value for money. The research councils took a bit of a risk on what I was going to do, and it paid off. The professorial fellowship ended up being an opportunity for me to develop methodologies, tools, theoretical frameworks, and multidisciplinary teams, and to build capacity within and outside the research community for policymakers and regulators.

I think you can create a direct path between that professorial fellowship and where I’ve ended up today.

Can you tell me a bit more about what the programme entailed?

Half of the project focused on citizens and political behaviour, while the other half focused on government and what government is doing with technology. What was exciting about this work was that I was working with big data and thinking about how it could transform social science back before the term ‘big data’ even existed. I was collecting data about political activity, such as petition signing and other ‘micro-acts’ of political participation, and a lot of exciting things developed out of that.

On the political behaviour side of things, I wrote a book with a multi-disciplinary team of co-authors called Political Turbulence, which focused on how social media shaped collective action. I ended up speaking about that book all over the world, because it was a new way of looking at politics. The central argument was that social media isn’t a source of all evil or all good, but rather it shapes political institutions and behaviour, and we need to understand the dynamics of that. It makes possible new ‘tiny acts of participation which can scale up rapidly and dramatically into large-scale mobilizations – or fail completely, which is what happens to most nascent attempts to do this. In this way, social media is a source of volatility and political pressure. In the years since we published that book, I think that line of thought has become kind of influential in raising the degree of attention that has been paid to certain kinds of political mobilisation and the importance of social media in that.

As part of this project, you did some work on how internet activity played an important part in revolutions and mass protests, with a focus on the Arab Spring. What sort of work did that entail, and how have you used the tools developed during that work since?

We analysed complex networks and large datasets, and we ended up developing tools in the field of computational social sciences, which is a field that had always been rather underdeveloped in the UK. I am now chairing some research projects which use similar techniques, but which focus more on the negative aspects, the threats such as hate speech and other online harms, that can come from these forms of participations.

You do have ‘good’ participation and ‘bad’ participation. Positive participation aims to change the world for the better through, for example, mobilisations and demonstrations. On the negative side of things, there’s conspiracy theories, the spreading of misinformation, extremism, radicalisation, and terrorism. However, no matter what you are measuring, you need the same tools – you need machine learning to detect and measure what is happening, and natural language processing to turn text into data.

Was there anything that really surprised you over the course of this research?

It was a real voyage of discovery, and I ended up working with people from disciplines I had never expected to collaborate with. The first two researchers I employed on the project were a computer scientist and a physicist, both of whom went on to faculty positions at the Oxford Internet Institute. So, this project showed me that multidisciplinary research of this could really happen, which is something which social scientists had always been rather dubious about.Ultimately, our research ended up being a forerunner within UK computational social science, which was a very positive development. The extent to which that whole field took off really surprised me.

How do you think your work has influenced policymaking?

Part of our project involved thinking about the possibilities of data science and data intensive technologies for government and policymaking, which is how I got involved in the Alan Turing Institute for Data Science and AI. I hope the stream of work that I have developed there, the Turing Public Policy Programme that I set up and now direct, has had a direct influence on the UK’s policy world’s understanding of data science.

And while traditionally, policymakers were not optimistic about the relationship between information technology and government, I think that attitude has shifted, and policymakers are now really interested. Through the Public Policy Programme, we are now working with over 70 government organisations, developing ways to improve policy-making, public services, government and regulation with data science and AI. We have, for example, produced the UK’s official guidance for the ethical use of AI in the public sector, the first comprehensive guide of its kind anywhere in the world. I don’t believe that programme would have happened without the research that I carried out in the ESRC progressorial fellowship, and I think it has had a really positive impact. In fact, the programme is being submitted by Oxford as an impact case study in the Research Excellence Framework.

One great piece for policymakers and researchers who are interested in learning more about my work would be a short article that I wrote last year with my Deputy Director, the economist Cosmina Dorobantu which was published in Nature called Rethink Government with AI. There is a lot of hype about artificial intelligence, but here we really explore what you can do with the more workaday tools and methodologies we have been talking about today – things like measurement, detection, and prediction. Those create both opportunities and threats for government, the public good, and public policymaking. Another useful data-intensive tool is agent computing, which can be used to simulate policy interventions and see what the effect might be without suffering potential unintended consequences. This can be a good way of dealing with uncertainty in advance. We are also building usable and theoretically rigorous frameworks for using these tools, based on values and principles such as fairness, accountability and transparency.

Photo by Thomas Jensen on Unsplash