Ethics and data protection in DIDYMOS-XR
The creation of large-scale digital twins comes with inherent ethical and privacy-related concerns, such as in the DIDYMOS-XR project, which seeks to capture and re-create scenes of public spaces. As the methods for creating the digital twins are still at an experimental stage, so too do the issues associated with the creation need to be researched.
To understand the extent of the ethical concerns, we will conduct a social and ethical impact assessment (SEIA) which is a systematic approach for identifying and assessing the risks associated with the project and providing a set of recommendations for mitigating the identified issues. This method makes it possible to consider and action the various concerns at different stages of development instead of simply documenting the risks. DIDYMOS-XR aims to create solutions that are ethical, privacy-aware and safe by design. Below we introduce some of the key ethical concerns related to creating digital twins:
Privacy:
As the digital twin will be created of a real-life scene, inherently there will be personal data of regular citizens captured as part of the initial scanning of the environment. Handling personal data immediately creates a data protection risk, which we aim to mitigate by anonymising the raw data from capture before further processing for point clouds etc. Algorithms used in DIDYMOS-XR are able to take the captured video and blur personally identifiable information such as people’s faces and license plates, making the creation and updating of the digital twin privacy aware by design.
Trust and transparency:
The use of AI, especially undisclosed use of AI, is topical in all parts of society, especially in research and innovation. For greater societal value and uptake of the project solutions, it is important to have the trust of the public and potential users of the product. One of the methods for achieving this is by being implicit in what the AI is being used for and how it may impact the user’s experience. We want to emphasise that any AI enabled solutions from the project can only support and inform decision-making and cannot make the decision for the user as the system may lack some context of the environment that the user is more aware of. To facilitate this, we would provide the application’s users with appropriate guidance or training to inform them of the benefits and limitations of using the system.
Social impact:
Digital twins have a potential to greatly impact and improve various aspects of our daily lives such as being able to include citizens in planning and city maintenance activities through interactive applications or by improving the safety of workers through being able to use digital twins for training purposes in otherwise dangerous environments. Hence, assessing the digital twins for their risks to safety is immediately important. As part of stakeholder consultations conducted in the project to understand the requirements potential end-users are expecting to see in our solutions, we found key concerns to be around the misuse of the digital twin, mishandling of personal data, and unrestricted access leading to targeting of vulnerable individuals. To mitigate data protection risks we have taken action to make data management practices compatible across the project partners and limit the access to personal data to only when strictly necessary. Additionally, all algorithms are rigorously tested in varied environments to ensure their robustness and reliability before system deployment.
Since DIDYMOS-XR is a collaborative research project, we discuss and agree the risks and mitigation measures with all the partners to establish shared responsibility of the issues and improve accountability. A great opportunity to discuss these issues was at our last consortium meeting which took place in Vilanova i la Geltru, Spain. We were hosted by project partners Neapolis on the first two days and Ficosa on the third day, giving us a great opportunity to see the pilot sites for our digital twin and the technology used for capturing the equipment.
Trilateral Research aims to use the findings from this research to inform future practical guidance for incorporating ethics by design in the creation of digital twins.
By Irma Poder
