Responsible innovations for intelligent intermittent lockdown

Jeroen van den Hoven

The drive to develop information technology tools capable of aiding in the fight against the COVID-19 pandemic has generated innovations in a short period of time. Can these innovations form the basis of responsible policy interventions? How do we know? We can benefit here from the lessons that have been learned in the past decade in the field of Responsible Innovation and Ethics by Design. This idea of Responsible Innovation (or in Dutch ‘Maatschappelijk Verantwoord Innoveren’) was pioneered by the Dutch Research Council NWO in the last decade. The idea was integrated into the Dutch Innovation Strategy of Key Sectors (‘Topsectoren’) and has by now gained worldwide acclaim. It was also used by the European Commission in its 80 billion Horizon2020 R&D programme to address ethical, legal and social aspects of technology and guarantee broad public acceptance and moral acceptability of new technology.

Lessons learned

The lessons learned from decades of ethics of technology are threefold and should be taken to heart in dealing with the pandemic: if we take our shared values serious we have to design for them and shape new technology in accordance with them (Design for Values). Second, in proposing innovations to solve urgent societal problems, we have to proceed responsibly and try to make moral progress and fulfil as many of our obligations as is feasible (Responsible Innovation). Finally, we need to cast our net wide, to include a great variety of disciplines, stakeholders, and a sufficiently generous systems perspective in which solutions are subsumed and without which we are not able to see clearly what they will engender in the bigger scheme of things (Comprehensive Engineering).

Design for Values

The sudden surge of technological solutions should reminds us of the fact that technology does not only deliver the advertised functionality. Certain technologies, architectures, applications, or services also favour, accommodate and promote particular ideals, normative conceptions of society, or preferred socio-economic models, whether this is done explicitly or surreptitiously, and whether it is intended or not.

Occasioned by the COVID-19 pandemic, our thinking and deciding in crisis and emergency mode, under conditions of deep uncertainty and incomplete information, only adds to the risk of obscuring the important value laden aspect of technology¹. This may not only lead to a distorted and flawed understanding of the values at play in our large scale social experimentations, but may also lead us to miss out on better options and mess up in terms of public acceptance and the undermining of trust in politicians and public health institutions. Furthermore, once we are out of crisis mode, it will compromise our ability to account for the values that were assumed and served by innovations².

The crisis context and high stakes of technological choices make it therefore especially important that particular values are made explicit, and that their implementations are carefully scrutinized and meticulously evaluated in practice. Furthermore, it is becoming clear that mere declarations of value commitments in this context are not sufficient^{3 4}. We need to tend to the coherence of our assumptions, expectations, predictions and beliefs, test the practical consistency of or moral and political judgements and evaluations, and systematically and transparently translate our shared values into design principles and technological requirements.
Value sensitive design methods support us in explicitly thinking of ethical considerations and moral values at early stages of development of technology in terms of requirement in design and research. This ensures that value discourse and moral discussions are not separated from what we actually do in practice in fighting the crisis, or from the socio-economic consequences⁵ of our choices. It ensures that our values are effectively applied, i.e. ‘functionally decomposed’ and operationalised, in the same way other high level and abstract requirements are functionally decomposed in engineering and design work.

Responsible Innovation

The successful implementation of the innovations that are necessary to deal with the intermittent unlocking, requires appreciation of genuine value conflicts and trade-offs that present themselves in the process. Value-sensitive design aims to go beyond mere declarations of value commitments and see moral values as non-functional requirements for which we can and ought to design. We must in particular avoid falling into the trap of false moral dilemmas and tragic choices dictated by technological determinism, market failures and private interests⁶. The main oppositions between health and the economy, between the economy and privacy, between privacy and accountability should not be accepted at face value⁷. Often, there are third options. Responsible innovation typically tries to transcend the dilemmatic character of these oppositions and encourages us to think of smart solutions, so we can avoid making a tragic choice that will make us regret our decision whatever we end up doing.⁸

Contact tracing apps

The implementation of apps for contact tracing⁹ presents us with choices between specific architectures, security models, and assumptions about users’ behaviour. We may consider for instance to provide individuals with the option to self-report their symptoms to other users of the app, or allow only patients that have been diagnosed to share their status once it has been verified by the medial authority. Even seemingly minor choices between different user interface designs, such as the choice of default settings, may nudge individuals into different behaviour causing value trade-offs¹⁰. We should also expect that seemingly minor technical design choices will cause completely unexpected problems, such interference of blue-tooth signal from a contact tracing app with the diabetes monitoring apps that connect to the subcutaneous sensor via Bluetooth¹¹. Safety, privacy, security, public health efficacy, user autonomy will thus all be affected differently by different combinations of technical features, and this need to be addressed proactively when we can still change things for the better.

Responsible innovation is the approach that aims to come up with new designs, concepts, and institutions or combinations of these which when implemented can expand the set of relevant feasible options for the solution of moral problems and avoid tragic or dilemmatic choices. Taken as an activity or process, responsible innovation enables moral agents to obtain relevant knowledge on the consequences of their actions and options, and evaluate them effectively in terms of relevant moral values. Responsible innovation thus differs from mere innovation or adding new functionality, as it attempts to add morally relevant functionality which creates new ‘third choices’ beyond dilemma’s and lead to new situations which allow us to do more good than we could have ever done before.

Comprehensive Engineering

By overly focusing on only one subset of socio-technical systems such as contact-tracing apps as a ‘silver bullet’ solution, we risk ignoring the wider systemic view, and the fact that the success of these apps also depends on the availability of appropriate medical and other infrastructures¹².

Comprehensive engineering then is the third key component in the thinking about and dealing with innovations that can aid engineers, developers, and providers of information technologies to adequately and responsibly respond to the global challenge of the current crisis. Adequate solutions to systemic problems such as a pandemic, are always systems solutions, which take into account technological aspects, human behaviour, values and norms. Comprehensive engineering is a form of complex systems engineering accommodating different aspects of socio-technical systems: dynamics, complexity, moral social and technical aspects. This is distinctively interdisciplinary and multi-disciplinary approach to the engineering, offering comprehensive analyses and suggestions for comprehensive solutions.

Systems thinking

To address the issues of fair distribution of risks, benefits and responsibilities decision makers need to be able to think comprehensively in terms of systems of systems in a sufficiently rich way. Challenges presented by the current crisis make it painfully obvious that ignoring just one component in this systems thinking can undermine even the most well-intended efforts. Going back to the example of contact tracing apps, we see for instance how far some of the proposed solutions that deliberately focus on specific system aspects such as convenience of data aggregation, may fail to achieve public trust and the sufficiently wide adoption prerequisite to their efficacy¹³.

Holistic view

Comprehensive engineering is not an approach that leverages understanding of social components to achieve successful deployment of technical systems. It takes a holistic view on the ‘systems of systems’ comprised of moral reason, institutions, procedures and individual humans with their own mental states who act in these contexts. As a research approach comprehensive engineering tries to understand how constraints and affordances of these normative structures interact with technical components, technical processes, and technical infrastructures. Engaging expertise of such disciplines as behavioural economics, evolutionary game theory, mathematics, experimental and applied ethics. As an engineering approach it aims at consilience, coherence, and collaboration in the design and re-design of socio-technical systems.