Principles
The ELSA Lab Defence works through two guiding principles: the Quadruple Helix model and iterative design. The Quadruple Helix model emphasises the importance of connecting stakeholders from different sectors. Through the iterative design process, we ensure that these stakeholders are involved in all stages of the research the ELSA Lab Defence undertakes.
Iterative Design
This picture shows the various cycles of the design methodology which is a key element within our lab. The ELSA Lab Defence will apply the design methodology based on a cyclic process of research, co-design, testing, analysing, and refining process and methodology. The different design stages and associated research work are closely related. The design cycle takes place every year (in total 4 times). This process is intended to improve the quality and functionality of the ELSA methodology.
Implement
The results from the entire process are disseminated for both the stakeholders and the general public. This dissemination involves for example the facilitating of talks and cafés open to the public, the sharing of lessons learned and outcomes within the NLAIC working groups, and the bringing together of expert groups to discuss findings and advises resulting from the ELSA Lab Defence. By setting the agenda and providing perspective for action, the outcomes provided by this ELSA Lab are more likely to be implemented and carried out by end-users, specialists, technicians, policy makers and directors.
Evaluation
The focus here is on examining whether the testing outcomes found prior are applicable in the field. If this is not the case, it is determined which part is due to methodology, and what is caused by organizational or other aspects an how can we adjust the ELSA methodology?
Test
In the testing phase, the ELSA methodology designed prior will be applied to specific use-cases and refined based on the experimental results. Experiments will be conducted both in the lab and in the field to test the implementation of the methodological framework. The experiments will be aimed at stakeholders from defence and policy, and they will focus on human decision-making and human-machine interaction in practice. The outcomes of the experiments will be made available by means of an advisory platform that will provide tailored advice and testing for ELSA-alignment of stakeholder’s AI projects.
Design
Outcomes from the research analysis in the various research lines (legal, ethical, design, public opinion and external use) are brought together to establish a ELSA methodology. The methodology provides design recommendations on how to improve a product’s alignment with ELSA values. It can also be used to assess how well products already align with current or anticipated ELSA values. The ELSA methodology will provide both guidance and technical instruments to implement and evaluate ELSA in military AI-based applications.
Research
In the research phase of the design process, researchers analyse both fundamental and applied scientific research on the legal, ethical, and social aspects of AI-based applications. Additionally, public opinion on the military use of AI systems will be collected and analysed to observe the societal impact of military AI applications. Finally, the adversaries’ use of AI will be monitored, along with the applicable international legal and normative frameworks.
Iterative Design
This picture shows the various cycles of the design methodology which is a key element within our lab. The ELSA Lab Defence will apply the design methodology based on a cyclic process of research, co-design, testing, analysing, and refining process and methodology. The different design stages and associated research work are closely related. The design cycle takes place every year (in total 4 times). This process is intended to improve the quality and functionality of the ELSA methodology.
Implement
The results from the entire process are disseminated for both the stakeholders and the general public. This dissemination involves for example the facilitating of talks and cafés open to the public, the sharing of lessons learned and outcomes within the NLAIC working groups, and the bringing together of expert groups to discuss findings and advises resulting from the ELSA Lab Defence. By setting the agenda and providing perspective for action, the outcomes provided by this ELSA Lab are more likely to be implemented and carried out by end-users, specialists, technicians, policy makers and directors.
Evaluation
The focus here is on examining whether the testing outcomes found prior are applicable in the field. If this is not the case, it is determined which part is due to methodology, and what is caused by organizational or other aspects an how can we adjust the ELSA methodology?
Test
In the testing phase, the ELSA methodology designed prior will be applied to specific use-cases and refined based on the experimental results. Experiments will be conducted both in the lab and in the field to test the implementation of the methodological framework. The experiments will be aimed at stakeholders from defence and policy, and they will focus on human decision-making and human-machine interaction in practice. The outcomes of the experiments will be made available by means of an advisory platform that will provide tailored advice and testing for ELSA-alignment of stakeholder’s AI projects.
Design
Outcomes from the research analysis in the various research lines (legal, ethical, design, public opinion and external use) are brought together to establish a ELSA methodology. The methodology provides design recommendations on how to improve a product’s alignment with ELSA values. It can also be used to assess how well products already align with current or anticipated ELSA values. The ELSA methodology will provide both guidance and technical instruments to implement and evaluate ELSA in military AI-based applications.
Research
In the research phase of the design process, researchers analyse both fundamental and applied scientific research on the legal, ethical, and social aspects of AI-based applications. Additionally, public opinion on the military use of AI systems will be collected and analysed to observe the societal impact of military AI applications. Finally, the adversaries’ use of AI will be monitored, along with the applicable international legal and normative frameworks.
QUADRUPLE HELIX
The Quadruple Helix model (Carayannis and Campbell, 2009) recognizes four actors in innovation: academia, government, industry, and society, each presented as a helix. Between them, there is a continuous dynamic push and pull relationship throughout the innovation process. The model was proposed to include the public and related sociological concepts such as media and cultural values in innovation processes and solutions for major social challenges. This should result in emerging knowledge and innovation ecosystems, well-configured for the knowledge economy and society.
To realise the Quadruple Helix model, ELSA Lab Defence works together with a consortium that includes parterns from each of these four actors. The partners, combined, have ample experience in (military) AI-systems, with a focus on legal (Leiden University, Asser), ethical (TU Delft), societal, (HHS) and technical (TNO, NLDA) aspects. Each consortium participant, on its own and in collaboration with the others, plays an important role for the success of the ELSA Lab Defence. Through this multidisciplinary ecosystem, ELSA Lab Defence shows how a theoretical model like the Quadruple Helix can be enacted in practice.