Based on the findings of this research work, we summarize five principles for guiding the design, development and use of digital public services: 1) respect for individual´s privacy, 2) effective data management and secure use of collected data, 3) explainability of automated decisions, 4) good purpose and equity, and 5) social consensus and legal enforcement. The proposed framework focuses on issues related to digital public services and is aligned to and extends other existing ethical frameworks, as those mentioned in Section 2.4. In particular, the first principle is directly linked to the privacy principle of Information Ethics. The second one directly concerns the quality and security of the collected data, which is partly related to the principles of privacy and access to information defined by Information Ethics. The third one, extends the responsibility principle defined by the COMEST Report on Robotics Ethics, while the fourth one is focused on not causing harm. Finally, the last one refers to issues related to good governance and legality, mostly present in all ethical frameworks.  The principles are explained below. 

Respect for individual´s privacy

New technologies allow public and non-public actors to pervasively and silently collect massive amount of information from individuals. To ensure the compliance with Principle R (human rights) of the R.O.A.M. Principles, data collected through digital services must respect individual´s privacy. In addition and despite that in some cases, the system collecting data requests the user to give consent, many persons do not have the capacity to understand neither the policy they are requested to sign, nor the consequences of the consent they give. Moreover, usually, they are totally unaware of the personal data that they are giving. Thus, on the one hand, public and private institutions should request consent for collecting personal data and other kind of data from users of digital services. Governments should apply their capacity and invest efforts to make sure that citizens understand the procedures used for collecting their data. Such efforts are aligned with satisfying Principle A (accessibility) of the R.O.A.M. Principles. On the other hand, individuals should have the right to know which data is being collected from them, for which purposes, and how such data is later used. In addition, open statistics published from collected data must ensure anonymity.  

Effective data management and secure use of collected data

Government-citizen and government-industry relationships are characterized by double mistrust – governments do not trust citizens and industry, and citizens and industry do not trust governments. On the government’s side, this is shown by an extensive amount of data requested to public service recipients. On the other side, citizens and industries are reluctant to openly give data to governments, to whom they pay taxes expecting to be well served. At the same time, they inadvertently give many relevant and sensible data to private companies whose ultimate aim is to get revenues from user’s data. Thus, ensuring effective data management, both by public and private institutions, could serve to build trust among societal actors. Particularly in the public sector, an effective data management will require adherence to commitments made by various stakeholders, i.e. proper data governance, and this refers to Principle M (multistakeholder approach) of the R.O.A.M. Principles.  Additionally and as part of proper ethics for data management, data should be used solely for the purpose for which they were collected. If further usage scenarios are conceived, the data collector should request citizen´s consent. In addition, governments need to ensure that data collected by them are secured either by them or by a third party. It means that collected data is not copied or altered by a third party, is only accessed by authorized actors, data is used correctly, and is destroyed when the original purpose for its collection ceases to exist.

Explainability of automated decisions

Automated algorithms, whether applying artificial intelligence techniques or not, implemented as part of information systems used to deliver public services should be transparent and accountable. It means, it must be possible to explain to third actors the criteria applied for making automated decisions, and different actors should be able to scrutinize software algorithms, if needed; i.e., the software code should be open for inspection, whenever requested.  In addition, data and data sources used by automated algorithms should be known. All this refers to open content, meaning that automated algorithms and their data, including data sources should be available, provided that existing legal frameworks, particularly considering intellectual property rights, as well as licensing restrictions are respected. The explainability of the automate decisions principle is related to the Principle O (Openness) of the R.O.A.M Principles. 

Good purpose and equity 

New technologies applied by public institutions as well as new innovative services delivered, either by public institutions or under their scope of control, should only serve good purposes and ensure that no harm is caused to individuals and society as a whole. In particular, digital services as well as data collected and used to deliver such services need to ensure that they do not cause physical harm to individuals, and do not lead to any kind of damage, including loss of credibility, social standing and fiduciary profile. In addition, the principle of equity must be ensured for digital public services, so that individuals have the same right to access and benefit from public services.  All such requirements are aligned with the Principle R (human rights) of the R.O.A.M. Principles.

Social consensus and legal enforcement

As it was explained, ethics depends on culture and at the lowest level, it depends on individuals. Thus, it has a key role in building consensus in society about core ethical principles for the new digital world. For that reason, governments do have the responsibility to seek citizens’ participation to broadly discuss issues that affect their daily lives and will affect their future. Likewise, they need to engage different social actors, like academia, industry, non-government organizations, and other key actors, for building consensus on  society’s value system regarding emerging technologies. All such efforts will require an effective multi-stakeholder governance. Thus, it is directly linked to the Principle M (multi-stakeholder governance) of the R.O.A.M. Principles. In addition, governments should ensure that such efforts are aligned and respect  international standards, like those defined as part of Internet Governance. Based on the consensus reached, the principles should be enforced by a legal framework defined to fulfill them.