On Wednesday 12 December, techUK’s second annual Digital Ethics Summit brought together experts from industry, government, academia and third sector to assess the progress made over the last twelve months to build the capacity and capabilities needed to recognise, identify and address digital ethical issues and concerns.
The Summit considered whether the practical action that has been taken is enough, whether the current public engagement on ethical issues is being successful and the progress being made to embed ethical decision making within organisations.
This sold out event was once again organised by techUK in partnership with the Royal Statistical Society, Wellcome Trust, Royal Society, British Academy, University of Oxford Data Ethics Lab, Open Data Institute, Digital Catapult, the RSA, The Alan Turing Institute, the Leverhulme Centre for the Future of Intelligence and The Nuffield Foundation’s Ada Lovelace Institute. It was sponsored by Microsoft, Gemserv, Intel, Taylor Vinters, Splunk, Kemp Little and QuantumBlack.
Opening the Summit, Antony Walker, Deputy CEO of techUK explained that one year on since techUK first held the Summit it’s time to take stock. He stressed the purpose of the day was not to debate whether we need ethics but “to agree actions and outcomes”. He encouraged everyone to consider what needs to happen to make 2019 the year we move digital ethics out of the conference room and into the board room and community halls across the UK.
Kate Rosenshine, Cloud Solutions Architect Manager, Data and AI Financial Services and Insurance from Microsoft then delivered a powerful opening keynote about the approach being taken by Microsoft to shaping ethical AI. She explained that AI brings great opportunity, but also great responsibility and the need to ensure AI systems are not developed in a vacuum but shaped by and for everyone. Kate highlighted the fundamental importance of educating the public so that they are asking the right questions moving forward.
In her address to the Summit, which can read in full here, the Digital Minister Margot James MP echoed the call for practical action. She highlighted the importance of building and retaining public trust so that technologies, such as AI, are working for society. The Minister outlined the Government's progress over the past twelve months, including the launch of the Digital Ethics Charter and the new Centre for Data Ethics and Innovation. She observed that digital ethics is increasingly becoming part of our daily conversations and is nowadays comparable in importance to medical ethics. Finally, she emphasised the need to continue to embed responsibility and ethics into the culture and fabric of companies and public institutions alike.
Following the Minister’s speech was a video address by the Information Commissioner Elizabeth Denham recorded two weeks before the Summit which can be watched in full below. In her speech the Information Commissioner highlighted how critical techUK’s 2017 Summit was for “driving the debate on data ethics in the UK”.
As well as providing an update on the work of the ICO over the last twelve months Elizabeth Denham explained that from her perspective data protection law has a role not just in legal compliance but also in ethics questions of “what is fair and what is transparency”. Reflecting that we are standing on the cusp of a “generational shift” in the concept of privacy she explained in detail that her current focus is on three issues; accountability, transparency and fairness.
Highlighting that digital ethics is a “complex space” Elizabeth Denham suggested the UK has an opportunity to lead the world with the UK’s competitive advantage coming from “strong laws including data protection” and a tech community that is “responsible and forward looking”. In closing the ICO explained that the “Democracy Disrupted” report, highlighted that the real harms identified in the case of electoral interference were “societal ethical harms”. She stressed that “democratic activities are too important to let technology get ahead of ethical thinking” and highlighted this as an issue she looks forward to exploring in more detail at techUK’s 2019 Summit.
Following the ICO, Sir Alan Wilson, Executive Chair of the Ada Lovelace Institute provided an update on the progress made since the creation and establishment of the Ada Lovelace Institute. This included the publication of the Institute’s prospectus, and the announcement of its first Board members. During his keynote address, Sir Alan provided a vision for healthcare of the future, drawing on a range of examples to illustrate that while machines can’t “think” they have the capacity to learn rules; creating huge potential but also raising ethical issues as consent, transparency and fairness. Sir Alan closed his speech by inviting industry to work collaboratively with the Ada Lovelace Institute over the coming months.
The Summit then broke into three morning sessions with keynote speakers and panels discussing issues including progress made over the last twelve months, public engagement and ethics in practice by business.
What progress has been made in the last twelve months and have we achieved what is needed?
At the first breakout of the day Sue Daley from techUK was joined by Roger Taylor, Chair, Centre for Data Ethics and Innovation, Tim Gardam, CEO, Nuffield Foundation, Rachel Coldicutt, CEO, Doteveryone, Hetan Shah, Executive Director, Royal Statistical Society and Lord Tim Clement Jones CBE, Chair the House of Lords AI Select Committee. Following opening remarks by Hetan Shah the panel discussed how all the new bodies and committees created so far should work together and what capabilities and capacity may still be needed to ensure that any gaps that exist now, or in the future, are identified and addressed. It was clear throughout the discussion that there is a need for the current ecosystem to be mapped and for bodies to be more joined up and collaborative in order to deliver success. It was felt that the focus should be not only on the ethical issues that need to be addressed but also how to empower people to create the change needed. In order to progress there is now a greater need than ever to get balance right between ethical and non-ethical use of AI and make a case of positive achievements to build public trust.
Public Engagement - What are the public telling us and how do we get this right?
The public engagement panel, chaired by Dr Stephen Cave, Executive Director of the Leverhulme Centre for the Future of Intelligence, included Sarah Castell, Head of Public Dialogue and Qualitative Methods, Ipsos MORI, Kate Coughlan, Head of Planning, BBC, Dr Kanta Dihal, Postdoctoral Research Associate, Leverhulme Centre for the Future of Intelligence, University of Cambridge, Hugh Milward, Director, Corporate, External and Legal Affairs, Microsoft and Jessica Montgomery, Senior Policy Adviser, The Royal Society.
This session discussed the findings from public research and engagement conducted on behalf of the Royal Society by Ipsos MORI. From the research it can be seen that the public’s hopes and fears threaten to distort their perceptions of new technology. The panel discussed how narratives of extreme fear can have potentially beneficial outcomes, for instance in building a safety culture, but are also more likely to cause misconceptions. To combat this, a range of credible scenarios can help navigate uncertainty in public debates. It was discussed how control lies at the centre of the public’s feelings of hope and fear and how being in control of tech encourages hope – the public tends to be fine with automation – but people fear this will be at the expense of the erosion of quality of life. This breakout session explored the notion that trust needs to be built by regulatory frameworks and guidance more so than narratives and concluded that the gap between tech adoption and societal intervention must be reduced. For more information on this topic, take a look at the Royal Society’s recent report on Portrayals and perceptions of AI and why they matter.
Ethics in Practice by Business
Following the keynote discussion, Riccardo was joined on stage by the panel’s chair Rob McCargow, Director of AI, PwC, Ivana Bartoletti, Head of Data Protection and Privacy, Gemserv, Matteo Berlucchi, CEO, Your.MD, Emma Wright, Partner, Kemp Little and Ana Perales, Strategic Transformation Director, Barclays.
Their discussion touched on the need for organisations to include the workforce in the AI debate and how human- AI augmentation will provide higher job satisfaction by supporting employees and eliminating mundane repetitive tasks from their workload. The panel felt companies should be transparent with appointments to their AI Governance Boards and ensure that employees are represented. Co-creating AI standards and principles with employees, will increase the chance of buy-in and having these standards operationalised through KPIs. Expanding the ethics discussion beyond its current bubble is key to make these issues more relatable and therefore acceptable to the public.
The morning panels were then followed by a keynote address by Professor Luciano Floridi who discussed ‘Socially good AI: challenges and principles’. Professor Floridi touched on five ethical principles for good AI, including: beneficence, non-maleficence, autonomy, justice and explicability. He argued the need for humans to move away from a human-centric model to more of an enriching planet-centric one, where AI can benefit people and society. He explained that digital ethics could be a long-term competitive advantage for the UK but cautioned that being risk-obsessed about technology could also have its disadvantages. He highlighted the importance of speaking about benefits and risks, not the other way around. His recent paper on this topic of AI for good can be found here.
After a busy networking lunch, the Summit moved back in to track sessions focused on ethics as a competitive advantage in an international context, data rights and controls, and ethics in practice in the public sector.
Ethics as a competitive advantage in an international context
This session was kicked off by Lenny Stein, Senior Vice President Global Affairs at Splunk. He provided a keynote address that stressed the importance of “data trust and ethical data conduct” to realising data driven technologies full potential and the need for ethics to be at the “core of every organisations culture”. Following his opening keynote Lenny Stein joined a panel discussion chaired by techUK’s Antony Walker along with Cecilia Bonefeld-Dahl, Director General at Digital Europe, Sarah Gold, Founder and CEO of Project by IF and Sana Khareghani, Deputy Director and Head of the UK Government’s Office for AI. The panel discussed how the UK’s ethics debate compares with other countries with speakers praising recent UK reports and the work of the ICO. Sana Khareghani commented that ethical discussions were happening outside the UK, focusing on common themes such as trust. Lenny Stein from Splunk suggested there is “less action” in the US on ethics and AI right now while Cecilia Bonefeld-Dahl explained the approach being taken in Brussels by the EU Commission on AI ethics. She then reminded the Summit about the importance of ensuring Europe’s younger generations are equipped with the skills to ask “the right questions” and suggested that the country who gets this right would have the competitive advantage. Sarah Gold also highlighted the need to remember the impact on people saying that ethical means “better for everyone” and explained how her organisation is helping organisations to put ethics into practice every day and what that means. Lenny Stein pointed out that ethics should be being seen by businesses as a “responsibility not a strategy”. On what advice the panel would give to SME’s on ethics Sana Khareghani stressed the importance of having “a voice in the conversation” in order to help shape the way ahead. In closing the panel Sarah Gold also highlighted that ethical questions are “shifting really quickly” highlighting the need to find ways to help people to understand and engage with the ethics debate.
How can we improve data rights and controls for individuals?
This panel session was chaired by Dr Claire Craig CBE, Chief Science Policy Officer, at the Royal Society and included the following panellists: Sophia Adams Bhatti, Director of Legal and Regulatory Policy, The Law Society, James Boyle, Senior Associate, Taylor Vinters, Darren Jones MP, Bristol, North West, Caroline Normand, Director of Policy, Consumer Association and Jeni Tennison, CEO, Open Data Institute (ODI).
The session began with the panel determining that ‘data ownership’ is an unhelpful term and that the individualistic notion of data ownership is misleading as data is both about the individual and the collective. The panel discussed that individuals often want some sort of agency over their data so they can determine how their data is used and shared. The panel agreed that data sharing, at the individual-level, is hugely complex and difficult in practice. James Boyle also highlighted the potential issue of user fatigue. Most individuals don’t have the time, knowledge or willingness to agree to every individual data transaction about them. The more realistic option is to help people to understand and determine their data rights and controls. Darren Jones MP emphasised that individuals want adequate protections and to know that the mechanisms in place are ‘fair’. The panel reminded the audience that data is already covered by a range of rights, for example, the right to protection and the right to privacy.
The panel recommended that we can improve data rights and controls for individuals by increasing the knowledge of the impact of the data use at the time of the data use, and through the governance of data in motion. This conversation needs to be led by Government at a macro-level but requires a multidisciplinary approach to ensure the public policy is transferable to mainstream debates. Sophia Adams Bhatti asked the audience to think about what the public policy response for the extraction or ‘mining’ of value from individual citizens should look like. The panel agreed that trust must be a broader part of the conversation. Caroline Normand, cited research from Which? showing that 89% of people were distrustful of companies mining and selling their data. We must therefore improve understanding of data usage and real-time impacts. Jeni Tennison also questioned how we can better reclaim and reuse data for public good.
For more information on this topic, take a look at the 'Data ownership, rights and controls: Reaching a common understanding’ discussion paper, which aims to encapsulate the rich and diverse discussion at a joint seminar between techUK, the British Academy and The Royal Society held on 3 October.
Ethics in Practice by Public Sector
This session began with a fireside chat between Shelley Campbell, Lead Product Owner, Intelligent Automation Garage, DWP and Mark Jennings, COO, Health and Public Service - Europe, Accenture. They discussed how DWP and Accenture are working together to deliver digitally-enabled public services. Mark Jennings stressed the importance of starting simple and thinking about consequences of action and the impact that introducing data-driven technologies will have on citizens and employees. Shelley Campbell also noted that she sees AI as a tool to help public sector workers to make the best possible decisions and empower people to dream big. In the panel discussion that followed Brhmie Balaram, Senior Researcher, RSA, Giles Herdale, Director, Herdale Consulting and Co-Chair of the Independent Digital Ethics Panel for Policing, Hilary Sutcliffe, Director Society Inside, Dr Indra Joshi, Digital Health and AI Clinical Lead, NHS England and Sam Roberts, Delivery Manager, Department for Digital, Culture, Media and Sport joined the discussion. The panel agreed that public trust and advocacy for a better public engagement is key. We also need to ensure that conversations on ethics are happening at a senior level. The importance of enabling people to have freedom of expression and to challenge decisions was also stressed. At the end of the discussion, the panel provided key advice for public sector organisations looking to embed an ethics by design approach, this included: encouraging organisations to try, test and learn from adopting data-driven technologies, showing a willingness to engage with the public and finally seeking opportunities for collaboration.
Plenary panel discussion - What do we need for 2019?
The final plenary session of the day discussed what needs to happen next in 2019 to take the digital ethics debate forward in the UK. Chaired by Antony Walker, the panel included Simon McDougall, Executive Director, ICO, Tabitha Goldstaub, Co-Founder, CognitionX and Chair, AI Council, Dr Natalie Banner, Wellcome Trust, Rachel Coldicutt, CEO, Doteveryone, Helen Mayhew, Partner & Chief Operating Officer, QuantumBlack, Richard Potter, Chief Technology Officer, Microsoft Enterprise Services and Cora Govett, Deputy Directors, Digital Charter, Digital and Tech Policy, Department for Digital, Culture, Media and Sport.
The wide-ranging panel discussion highlighted a number of common threads and themes from the day. As well as highlighting that the ethical debate needs more diversity of perspective – with Natalie Banner asking the audience to consider who’s missing in the room today? – the panel stressed the need for a much broader church to be involved in the ethics debate. Tabitha Goldstaub raised the point that we need to demonstrate why we are developing AI systems; showcasing the positive impact that this technology is having on people’s lives right now. The importance of thinking about how to engage effectively at the citizen-level was a common theme of the discussion with the point made that people should be able to not care about these issues and still be safe.
Rachel Coldicott highlighted the need to engage with wider society to prevent the AI ethics community from becoming too insular or disconnected from the societal changes happening all around us. This is an issue to be explored further at the doteveryone’s Responsible Tech 2019: The New Normal event on Thursday 31 January.
On the question of what needs to happen in 2019 Cora Govett suggested that the debate should be on where ethical boundaries are while Helen Mayhew agreed that we need to be discussing actions next year not just engaging in a philosophical debate. Simon McDougall pointed out the need to meaningfully engage with the sectors that are rapidly becoming data-rich, but don’t necessarily have experience in this area. He also called for action on some of the ideas discussed during the Summit through genuine collaboration. While the panel felt progress has been made this year it was stressed that technology has also moved very quickly in the past year.
Bringing the Summit to a close, Antony Walker thanked all the speakers, partners and sponsors for making this year’s Summit a huge success. He highlighted the sheer number of different issues that have been covered this year during the sessions, in the keynotes and also during the networks breaks.
Looking head into 2019 Antony highlighted techUK’s ongoing commitment to working on digital ethics issues and looked forward to welcoming everyone back in December 2019 to what has now become an annual event for techUK. Antony closed by stressing that next year’s Summit will aim to not just assess the progress made during the year ahead and to test the homework of the people in the room, but also to present and share real examples of ethics in action from 2019.
If you missed any of the action from the day, keep an eye out for the speaker slides and panel videos, which we’ll be uploading later this week. For more information on techUK’s digital ethics work programme, please do not hesitate to get in touch with Sue Daley or Katherine Mayes. Finally, I’d like to encourage everyone to continue this important conversation on Twitter using the #AIEthics.