By Sir Nigel Shadbolt, Co-Founder, Open Data Institute
Seven years ago Sir Tim Berners-Lee and I founded the Open Data Institute (ODI). I’m often asked whether we’ve achieved the things we set out to do in 2012. A lot has been accomplished, but a lot has changed in that time. There’s no doubt that we have plenty of unfinished business; as an organisation, as individuals and as wider society. It’s unclear whether we will ever be able to declare success in a world in which data is becoming more ubiquitous, more varied and more important; where urgent issues are constantly arising.
For example, we hear a lot more now about the importance of ethics in the collection and use of data. Across the world there are dozens, if not hundreds, of ethical codes and frameworks for data and Artificial Intelligence (AI). What’s not clear is how much of this is ‘virtue signalling’ or if countries and organisations are prepared to act in support of ethical principles. A survey out this week by YouGov (for the ODI) shows that whilst 87% of the public think it’s important for organisations to use personal data ethically, most are unconvinced that they will. Only 31% of those surveyed said they trust government, whilst for social media organisations and marketing and advertising agencies, the figures were just 5% and 3% respectively. The full dataset can be viewed here under an open licence.
The ODI, along with others, is working hard to develop our understanding and practice around data ethics and there is precedence from other fields. Medical ethics guides and informs healthcare professionals, clinicians and researchers; it’s hard to imagine a time when it wasn’t so. But medical ethics took decades to develop. It took careful thought, research and determined effort. In data and AI, I see a lot of well-meaning initiatives and commentary but this alone is not enough. It needs deep research that goes beyond technologists, and involves, for example, social scientists, psychologists and philosophers.
That’s precisely why we are setting up the Institute for Ethics in AI at the University of Oxford, to study the ethical implications of AI and advances in computing technologies; rooting this activity in the Humanities within the Faculty of Philosophy. In the meantime, the ODI has provided practical help so that people can find their way through the various ethical challenges that arise in our data driven society. Their Data Ethics Canvas has been widely used and recommended by organisations including Co-op Digital, the BBC’s ethics working group and Doteveryone.
Alongside ethics, access to data is still a challenge and a barrier to innovation. Data is a new form of infrastructure, something we now rely on in our societies and economies, and in our daily lives. Open data, which anyone can access, use and share, is the foundation of that data infrastructure. We have made real progress since 2009 when I first became involved in open data. There have been significant releases across a wide range of domains from transport to the environment, health to policing, legislation to geospatial data. But important parts of our public data infrastructure remain problematic. For example, access to open address data remains elusive. Even the data that underpins our electoral process remains unavailable from official sources as open data; data like the locations of polling stations and lists of candidates standing for election. Instead, this data it is maintained by organisations like Democracy Club, and sustained by voluntary effort.
We’re trying to create a system that the public, private and third sectors can thrive on; a data infrastructure to match our physical infrastructure. An infrastructure that needs to be designed, built and maintained; and that should be at the heart of our national data strategy. We have started to see more commercial organisations opening up access to their data, and participating in the development of this infrastructure but concerns about security and commercial sensitivities make for slow progress.
Despite these various challenges new models and processes for widening access, and sharing data safely are starting to emerge. Earlier this year, the ODI ran three pilot ‘data trusts’ with the Office of AI. Data trusts could be game changers, especially for the private sector, in terms of their approach to data sharing. The ODI defines ‘data trusts as legal structures that provide independent, third-part stewardship of data.’ In short, data trusts underwrite, or mitigate, some of the practical and ethical risks attached to data sharing.
So as we celebrate the dawn of our eighth year at the ODI, I believe our work remains just as relevant and important. The burgeoning data ecosystem, the complexities of technology and the challenge of ethics keep us busy; there is plenty of unfinished business.
-----
Professor Sir Nigel Shadbolt is the Executive Chairman and co-founder (with Sir Tim Berners-Lee) of the Open Data Institute. He is Principal of Jesus College, Oxford, a Professor in the Department of Computer Science at the University of Oxford and the co-author of The Digital Ape: How to live (in peace) with smart machines (2018)
This opinion piece was first published in NS Tech on 13 November 2019: Sir Nigel Shadbolt on privacy, the ODI at seven and ethical AI virtue signalling