The validation methodology used to tailor the RDT Framework to the specific context of application

An overview on the methodology used for validating the Responsible Disruptive Technology (RDT) framework and tailoring it to the specific context of application: the case of the public organization AI-technology for misinformation handling.

The objective of the ETAPAS Project is to improve public service delivery  for  citizens  and  the efficiency  of  public  sector organisations,  by  facilitating  the  ethical  adoption  of  Disruptive Technologies (DTs) by public sector organisations. One of the main outcomes of the first project phase was the development of the Responsible Disruptive Technologies (RDT) framework, whose scope is to provide a tool for the assessment and governance of the ethical, social and legal risks of disruptive technologies (DT)-based applications. The framework includes a generic Code of Conduct, a Legal framework based on the European ones, a set of ethical risks and social impacts of DTs and the RDT indicators aimed at measure these risks. Given the importance of factors such as the type of service provided, the stakeholders involved (both internal and external), the type of technology adopted, and the frequency of use of the service for the identification of the ethical risks and social impacts stemming from the adoption process, a validation methodology has been developed to help PAs in analysing the context and tailoring the RDT Framework to the specific context of application.  Specifically, the process foresees (1) an Initiation Phase where the multidisciplinary team is set up to conduct the assessment and the context of application is analysed; (2) an Assessment Phase in which principles and risk categories are prioritised on the basis of the application context and the actual assessment is undertaken; and, finally, (3) a Final Validation & Fine-Tuning Phase to ensure the DTA reflects the adopters’ needs and also addresses the ethical risks identified. This process has been recently applied in the fourth ETAPAS use case which include the deployment of AI for fake news detection and prioritization of emerging issues in the municipality of Katerini (Greece). The application of the process inform both the refinement  of  the  conceptual RDT  framework and  the  development  of  the ETAPAS prototypical platform supporting the assessment of the DTAs. At the same time, these results contribute to the development a Governance model for the PAs that intend to adopt DTs, which could be used as such or adapted to the organisation’s context and applied to the various use cases (based on PA’s in internal decisions).