On May 1st 2018 the SHERPA project was officially launched. In this exciting and prestigious international project, different partners come together to investigate, analyse and synthesize our understanding of the ways in which Smart Information Systems (SISs), involving big data analytics and artificial intelligence, impact ethics and human rights issues. Prof. dr. Philip Brey, Dr. Kevin Macnish and Postdoctoral researcher Dr. Mark Ryan of UT are one of the main partners of the project coordinated by Prof. Bernd Stahl from de Montfort University, United Kingdom. The other partners are the European Network of Research Ethics Committees (EUREC), the University of Central Lancashire Cyprus (UCLAN Cyprus), Trilateral Research, Nederlands Normalisatie-Instituut NEN, Mutual Shoots, AEQUITAS, F-Secure, and the European Business Summit. SHERPA, short for ‘Shaping the ethical dimensions of information technologies – a European perspective’ is funded through the European Union Horizon 2020 “Science with and for Society” and will end on November 1st 2021.
The new potential offered by Big Data and Smart Information Systems (SISs)
SISs, processing vast amounts of data, allow us to identify correlations that humans would never be able to find that can consequently inform our decisions and practises, offering enormous benefits to society. To illustrate what benefits SISs might offer to society, one can think of the collaboration between the Royal Free Trust and Google DeepMind. In this partnership, the Royal Free Trust asked DeepMind to create the mobile application STREAMS to timely predict acute kidney injury (AKI) of patients. Each year about 40,000 individuals die from kidney failure and many of these deaths could be prevented if we had software capable of detecting informative patterns in the vast amounts of data. SISs not only offer new invaluable information, they provide additional benefits such as greater speed, accuracy, safety, or reduced costs.
Protecting citizens and society from risks
However, to reap these benefits we must better understand and carefully implement these systems. Kevin Macnish says:
“I think that society is on the cusp of integrating AI and big data into a myriad of different areas and industries over the next 10 years. To do this without adequate forethought would involve tremendous amounts of risk being imposed on the public without their consent. Instead, we have a moral duty to investigate and thoroughly understand the ethical challenges presented by these technologies both now and looking ahead into the future”.
The public’s reaction to the partnership between the Royal Free Trust and DeepMind illustrates that addressing these implications is not an easy task. A lack of consent on the principle of ‘direct care’ was ruled faulty, public trust was undermined as a consequence of the delayed announcement of the collaboration, no clear explanations were provided to the patients regarding the necessity of such extensive and personal data, and the partnership was criticised for not approaching relevant public review bodies. As such, this collaboration to use a SIS for societal benefit illustrates that good intent is not sufficient to reap desired benefits. Since the public outcry, the Royal Free Trust and DeepMind have taken several steps to regain trust from the public and patients. Moreover, they have acknowledged the need for appropriate guidance in such collaborations dealing with SIS. Mark Ryan adds:
“Artificial intelligence and algorithms are deciding factors that are having drastic effects on our values and way of life. The importance of the SHERPA Project is to address the ethical and social implications of these technologies in our lives and how our relationship and dependency to them will evolve in the future”.
Improving the public’s health with Smart Information Systems
During this project, new ways to understand and address the challenges brought forward by SIS will be developed. Kevin Macnish says:
“I have come across projects in AI and projects in big data, but none that have yet combined the two on this scale. Furthermore, this project puts ethics at the heart of the research, which is crucial as we seek to understand the potential impact of these technologies”.
The SHERPA project has more unique features. There will be an ongoing and close collaboration with a stakeholder board. This stakeholder advisory board includes, amongst others, the European Data Protection Supervisor, Facebook, Google, the European Organisation for Security, Big Data Alliance, Data Value Association, Postbank Germany, Helsinki Foundation for Human Rights, and the European Economic and Social Committee. With six stages in total, the first will be to conduct 15 extensive case studies, such as the collaboration between DeepMind and the Royal Free Trust, and scenario studies to visualise how SISs can introduce ethical issues and how they might affect human rights issues. Moreover, these will be conducted together with organisations that develop and use such systems. During the second stage the partners will collaborate with the stakeholders to discover exactly which concerns they have with SISs and which solutions are realistic in practice, manageable and therefore, preferable. In the third stage, the project members will create and publish a workbook describing responsible development of SISs. In stage four, technical and regulatory options will be presented. Stage five serves to validate and prioritise the various proposals through multi-stakeholder focus groups. The final stage will be to advocate these results, and implement the most promising solutions through targeted dissemination and communication activities. This unique project will thus offer greater understanding and in turn benefit the further integration of SISs in society. Philip Brey says:
“We develop practical tools for better development of these technologies and better policies, including ethical guidelines, research ethics protocols and recommendations for law and policy. The close collaboration with key stakeholders ensures that we do not just do solid scientific research, but also develop practical tools making it likely that our results will be endorsed and used”.