Social validation of multidisciplinary design in ICT and architecture

Abstract

This paper brings the case of research centres operating at the crossing of the fields of architecture and ICT (Information and Communication Technologies). These centres aim at innovative multidisciplinary research on, for instance, smart buildings, for producing design projects that address societal issues. Yet they lack proper methods for validating these projects for their social impact. We discuss validation practices in ICT and architecture, demonstrate their insufficiency for social validation, and explore methods that can provide this validation. In software development, it is common to do validation tests of developed products. These include performance testing where applications are tested for their stability and response time, and user acceptance testing where client/customer test applications with real-time scenarios. Assessing how the application may impact society beyond the primary users is rarely done. Examples of social failures in ICT are loss of privacy and bias in AI algorithms. In architecture validation is mandatory by law for some parts of the design, such as load-bearing structures, safety against fire and energy efficiency. While design projects are assessed against such construction regulation, assessment of acceptance of buildings by future users is only taking place for a minority of designs. Assessment of how architecture can impact society in the future is seldom done. The strongest example of societal failures is social housing after World War II, such as on the outskirt of French cities. So, for multidisciplinary research that combines ICT and architecture for addressing societal issues, one has to look beyond the existing validation practices. Firstly, in architecture there are a few instruments for forecasting likely effects or assessing the impact of architectural on users (e.g. post-occupancy evaluation [3], and space syntax [2] that analyses how people move through buildings). Architects and urban planners rarely use these instruments and, if they do, apply them after construction, which might lead to substantive costs (if reconstruction or demolishing is needed, as for, e.g., post-war social housing). A step forward would be to use these instruments before construction. Secondly, the European Commission has recently introduced new requirements in research project proposals to explain the potential impact of technological development when AI is involved (see, e.g. [1]). This is a step towards identifying potential social negative impact, but it is still not enough since there are no clear instruments available for giving these explanations. A third route towards finding validation methods is to look at current developments in engineering and design to realise moral and social values in the design of technology and products (e.g. [4]). This development includes finding methods for validating these values, which may be transposed when the values involved in multidisciplinary ICT-architecture design projects are explicitly articulated.