Interoperability and (Data) Ethics in the artificial intelligence era: an important principle to be included in interoperability frameworks
Interoperability and (Data) Ethics
The OECD document Good Practice Principles for Data Ethics in the Public Sector highlights the importance of data ethics in the public sector, including 10 good practice principles and specific actions which can support their implementation. Integrity and management, arrangements and trustworthy data access, data ethical considerations in governmental organizations, control over data in processes, publication of source code, open data, accountability or risks management, are some of the elements included.
Nevertheless, the document surprisingly does not include explicit mention to interoperability, despite the fact that interoperability is considered by literature and practice as one of the key core elements in digital government. In a moment where digital transformation is so critical in the public sector, this would have been a very good opportunity for OECD to explicitly reinforce this point, evangelizing and spreading the word on this important issue.
At the same time that interoperability puts the focus on data and information sharing, data is what is used to train models and, eventually, used in the artificial intelligence (AI) systems. As AI systems learn from data and data sources are diverse, the data shared could also be collected and used to feed AI systems. The range of implications of data-related variables in this kind of systems are enormous and can go from the bias or important variations in outputs and results, until the discovery of new -and unexpected- information from the data analyzed, to cite some of them.
Ethics has become a required component when talking on technology. For those that still have doubts about that, I suggest you to visit the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems, where you can find resources -including the family of the IEEE P7000 standards- on transparency of autonomous systems, personal data AI agent, or algorithmic bias considerations, to cite some of them. Ethics is at the core of this new generation of standards.
As an interoperability principle can be considered as "fundamental behavioral aspects to drive interoperability actions" -I follow here the concept included in the last version of the European Interoperability Framework-, at this point, the link between interoperability and ethics is definitely a must to be taken into account in the digital government context. We have reached a point where ethics and the ethical use of data should be urgently adopted in policies and included within the interoperability frameworks, as a general principle to be followed.