I joined the department of AI at the University of Edinburgh in 1978 as a PhD student and then went on to establish the AI Group at the Department of Psychology, at the University Nottingham in 1983. It was here that I led the first commercial spin-off, a rule based system for designing industrial plants.
In 2000, I moved to Southampton’s School of Electronics & Computer Science to research the next generation of the World Wide Web and co-founded Garlik, which won the BT IT Flagship Award and was subsequently acquired by Experian.
In 2009 I was appointed, with Sir Tim Berners-Lee, an information advisor to the UK Government. This work led to tens of thousands of non-personal data-sets being made available as open data. In 2012, I co-founded the Open Data Institute with Sir Tim and three years later became a Professor at the University of Oxford and Principal of Jesus College; I was recently announced as the lead in setting up Oxford’s Institute for Ethics in AI.
In addition, I have also published hundreds of research articles and am the co-author of The Spy in the Coffee Machine and The Digital Ape. I am a Fellow of the Royal Society, and of the Royal Academy of Engineering, as well as being a Fellow and former President of the British Computer Society. I was knighted in 2013 for services to science and engineering.
Launching the ODI has to be one of them; it has helped organisations understand the power of sharing data and the benefits of using open standards. We have trained thousands of individuals and helped to raise awareness around how to build a healthy and sustainable data ecosystem; recognising that data comes in all shapes and sizes with different requirements to keep it open, shared or private. We work on developing policy around new kinds of data sharing arrangements, such as data trusts and cooperatives.
No one individual – there are lots of role models out there for different qualities at different times.
We probably reached peak AI for this time round the technology adoption cycle. Periodically, since the 70s, AI has been held up to be a major disruptor in the digital space. Each time advances are made; each time a degree of reassessment follows.
It turns out to be harder to automate many of the things humans do than we imagined. Building blended teams of human and computational capabilities is usually the way forward. In particular sectors, we have started to see the benefits of open standards for data interoperability – open banking is a notable example.
I see the continued increase in the “Vs” of our industry – volume, variety and velocity and an increasing concern over veracity and validity. Can we trust the data to be what it claims to be and how representative is it of the problems we are looking to address? I also expect an increasing focus on regulatory, governance and ethical frameworks, which will be essential to allow continued access to, and analysis of, existing and new forms of data.
Predictive analytics across whole swathes of society. Health and well-being is set to be transformed as we understand how to predict and anticipate all sorts of outcomes from an ever growing range of data. The same dynamics are at play across many different sectors; transport, energy, finance, retail, leisure etc. With these data driven insights we face the challenge of appropriate governance. We need to ensure that access to these predictive insights is equitable and fair, is able to be held accountable, used proportionately and so on. We are back to ethics.
The biggest challenge is often not technical – but cultural and organisational. Organisations rush to embrace a particular technical solution without appreciating the skills and training required to take advantage of the solutions on offer. Or else stakeholders are insufficiently involved in the requirements elicitation in the first place. What sort of digital transformation, for what purpose, to what end? Technical challenges can’t be ignored and ensuring that the underlying data engineering has been done so as to enable integration of data assets and resources is crucial. There is still huge friction in a lack of data interoperability.