Abstract

The rapid evolution of the technological landscape, characterized by advancements in Generative AI, Cybersecurity, IoT, and Cloud Computing, presents both opportunities and challenges for official statistics. Eurostat recognizes the imperative to stay abreast of these trends to remain relevant in a digital era. The EU Artificial Intelligence Act (2024) defines AI systems as autonomous, adaptive entities capable of influencing environments through predictions, content, and decisions. This contribution will explore the necessity of AI in official statistics, driven by the fear of missing out on technological advancements and the need for thoughtful consideration of appropriate AI use cases.
The presentation will draw on potential applications of AI in data collection, editing, dissemination, and processing, highlighting examples under exploration in Eurostat such as automatic classification for web intelligence systems, chatbot advisors for users of open data, and machine learning for nowcasting economic indicators.
Example of collaborative undertaking at different levels such as the European Commission community of practice and European Statistical System One-Stop-Shop on AI/ML will tend to demonstrate the value of cooperative approach for model customization, expertise exchange, and resource pooling within the statistical community. The paper will also explore the way ahead for maintaining statistical community ahead of the wave of technological revolution as stake.
However, the integration of AI also raises ethical and regulatory concerns. The presentation will address the need for adapting responsible AI principles in official statistics, ensuring that AI applications are for the public good, fair, unbiased, transparent, accountable, safe, and secure. We will present the development of AI ethical guidelines, risk management strategies, change management approaches and compliance with legal frameworks in the context of Eurostat.
The impact of development AI on official statistics, including the risks of unreliability and increased public mistrust and disclosure of confidential statistical information will be examined. Generative AI's role in changing information access dynamics will be a focal point, emphasizing the need for responsible implementation and communication.
In conclusion, the presentation will highlight the importance of mastering AI technology to harness its benefits and mitigate its risks, ensuring a future for Eurostat and the European Statistical System in the age of AI.