The eGovCollab project involves the University of Macedonia (UOM) and the University of SouthEastern Norway (USN), with additional support from the Greek National Infrastructure for Research and Technology (GRNET). The project aims to strengthen the collaboration between the two universities by exchanging expertise in fields like e-government, AI, and project management, as well as by forming long term cooperation strategies. Furthermore, it promotes the exchange of best practices between Greece and Norway, supported by GRNET, to enhance the delivery of public service personalized information. A Memorandum of Understanding was signed formalizing this partnership, and practical policy recommendations were developed. This section presents the approach we followed to complete each aim of the project. Step 1. Literature review and meetings. In this step, we conducted a review of the available literature to identify research papers that focus on providing personalized information in eGovernment. Furthermore, we conducted a series of meetings and a workshop to identify and exchange best practices and ideas between the two participating countries. Step 2. Proof of concept. In this step, following the Design Thinking Process Guide [1], we developed and tested a proof-of-concept AI-enabled tool for public service personalized information provision [2]. Step 3. Policy Recommendations. In this step, based on the findings of the literature review, the meetings, the workshop, and the experience gained through the project, we derive a series of policy recommendations.
2. Implementation Approach 3. Policy Recommendations In this section, we list one of the main outcomes of the project, i.e., the policy recommendations (Table 1). Problem
How it helps 1 Disorientation of citizens through vague and general
information 2 Risks to privacy due to the use of personal data 3 Digital exclusion of vulnerable groups 4 5 6 7 8
inconsistent information flows High cost of using AI models
Difficulty in quality control of mass-produced
content Lack of evaluation of personalized information systems
Adopt a citizen-centric approach and provide personalized information using appropriate applications Use anonymization and data generalization techniques
interaction between citizens and
services
It protects citizens' privacy while maintaining the usability of the data
Provide educational material It ensures universal access and through alternative channels such enhances social inclusion
as SMS and IVR
Complexity and Adopt standardized data models It enhances interoperability, heterogeneity of data (e.g., CPSV) and use technologies, enables personalized information and such as knowledge graphs increase the accuracy of responses
Develop proactive services It increases citizen satisfaction by using AI and machine learning providing them with services they are
entitled automatically Explore open-source solutions It reduces costs and increases the
or hybrid models sustainability of services Develop semi-automated It ensures consistency and quality control tools accuracy without excessive cost
acceptance evaluation methods (e.g., SUS, TAM)
It provides evidence-based data for the continuous improvement of the systems This research was funded by the project “eGovCollab: Strengthening research and know-how transfer for good electronic governance”, MIS Code 5223791, benefits from a 54070.39 funding from the EEA Grants (funding period 2014-2021), which represent the contribution of Iceland, Liechtenstein, and Norway towards a green, competitive and inclusive Europe.
Declaration on Generative AI The authors have not employed any Generative AI tools.