A call to avoid citizens being outwitted, devalued and marginalised by AI innovation

A response to the Artificial Intelligence and Robotics: innovation, funding and policy priorities event hosted by the Westminster eForum by Dr Matt Roach, Computer Science Lecturer at Swansea University and Senior Development Officer for the Computational Foundry, inspired by many in depth discussions with Colleagues

In a session chaired by Lord Clement Jones, Chairman of Lords Select Committee on Artificial Intelligence, we heard about the capabilities and potential for AI to enhance our lives and grow the economy. Cisco’s Director of Government and Corporate Affairs, Matt Houlihan reports that reports of global revenue for AI over the next 5-10 years vary widely, but estimates of $100B are not uncommon. A highlight was Dame Wendy Hall’s presentation of a report (co-written by Jérôme Pesenti, VP of AI at FaceBook) named: “Growing the artificial intelligence industry in the UK”. In the report, it states: “It has been estimated that AI could add an additional USD $814 billion (£630bn) to the UK economy by 2035, increasing the annual growth rate of GVA from 2.5 to 3.9%.”

We also heard from Prof Philip Nelson, Chief Executive of EPSRC, of the Government’s increasing investments in research and spoke specifically of initiatives in priority areas of Enabling Intelligence and Applications of AI, and the pressing national need to train researchers who can provide leadership in these future keystone technologies. Furthermore, in response to questions from the audience, various panel members gave examples of AI having a positive impact in a range of key sectors. Indeed, in our work we have seen this healthcare, manufacturing and digital services.

However, a session chaired by Rt Rev the Lord Bishop of Oxford demonstrated that you do not need to look far for evidence that highlights widespread and growing concerns around the coming transformations promised by big data and artificial intelligence. These concerns were voiced by other panel members throughout the day, including Dr Rannia Leontaridi and Gila Sacks, directors of BEIS and CDMS respectively. They are not alone. In July 2017, perhaps the most famous tech entrepreneur of our day, Elon Musk, said, “AI is a fundamental existential risk for human civilization, and I don’t think people fully appreciate that, [ . . . ] (AI) is the scariest problem.” More prosaically, in its January 2018 review, Deloitte posits that, “. . . at a business level, large “big data” and AI projects often fail to deliver,” and poses the rhetorical question, “What’s crucial? Ensuring it’s designed to help humans think better.” The review concludes with a call to centre innovations in enhancing human machine collaborations.

How we can respond to this clear and challenging need? Scientists at the Computational Foundry believe establishing an approach for human-centred and shaped approaches to data driven and intelligent systems is part of the answer. Yes, there will be many — and essential — initiatives globally where computational researchers innovate in terms of fundamental data analytics and AI algorithms and processes, covered by Dame Wendy Hall’s talk today and detailed in her review on growing the artificial intelligence industry in the UK. My colleagues and I would like to see an approach that would lead breakthroughs in core aspects of computational science represented in multidisciplinary research teams that can enhance the human experience of data and intelligence. A focus on nurturing an environment where researchers work to ensure that such systems are trustable, understandable and negotiable by and with humans.

We need to lead a change in industry and academe by having deep technical skills and knowledge in core computational science that are honed with a simple but profoundly important question. That question is: what breakthroughs in core computational science are needed to enable people to feel in central control of their futures rapidly being transformed by advances in data and intelligence driven technologies?  Moreover, this requires responsible innovation, embedded at the heart of every research question. This viewpoint will always see the human as a first-class citizen in the future physical-digital world, not outwitted, devalued or marginalised by the expanding capabilities of machine computation, automation and communication.

While it might be assumed by others that human-centred thinking for data and intelligence is the special remit of Computer Science’s Human Computer Interaction sub-discipline, we need to be radically more adventurous and ambitious than this. We need to have multidisciplinary, diverse computational scientists across a spectrum of theoretical-to-experimental topics so that all can contribute to the human enhancing overarching agenda of economic and social improvement through the adoption of AI. By way of example I pose some sample research questions:

  • How can we use visualisation and logical reasoning to explain to a user the decisions made by an intelligent machine?
  • How can a user negotiate, through new forms of interface with such a machine to shape and alter decisions and behaviour?
  • What is needed in terms of formal specification, algorithmic and interaction terms to enable people to trust these new forms of technology?

I would like to send my thanks to the contributors and organisers of this event, and it was a delight to be part of a room full of highly motivated people to make a positive difference to the world, to understand and deal with the complex challenges and exciting opportunities this topic presents.