Materials in order of publication, from newest to oldest
- DPI-AI Framework - Vision paper on Building AI-Ready Nations through Digital Public InfrastructureOpens in a new window. (Center for Digital Public Infrastructure, January 2026)
Vision for the application of artificial intelligence at the organizational system and ICT system levels as part of the digital infrastructure of public administration. - Generative AI and the EUDPR. Orientations for ensuring data protection compliance when using Generative AI systemsOpens in a new window.. (European Data Protection Supervisor, 28.10.2025)
The revised guidance offers clearer and more practical instructions for the responsible development and deployment of generative AI tools. - Responsible AI: A Strategic Guide to Address the Elephant in the RoomOpens in a new window. (Responsible AI Institute, August 2025)
- The New Politics of AI: Why Fast Technological Change Requires Bold Policy TargetsOpens in a new window. (Carsten Jung, Institute for Public Policy Research, February 2025)
"Merely accelerating AI deployment and hoping it will deliver public value will likely be
insufficient. We argue that AI needs to be directed towards societies’ goals, via
‘mission-based policies’. How these are determined and measured is at the heart
of this new politics." - Ethical and regulatory challenges of AI technologies in healthcareOpens in a new window. (National Research Council of Italy (CNR), 26.2.2025)
The research article addresses important ethical and legal issues related to the implementation of AI systems in clinical practice. - Ohjeistus generatiivisen tekoälyn hyödyntämisestä työn tukena ja apuvälineenä julkisessa hallinnossaOpens in a new window. (Ministry of Finance, 26.2.2025, in Finnish)
- Experiments using generative AI in public administrationOpens in a new window. (Publications of the Ministry of Finance 2024:48)
This report compares experiments using artificial intelligence in public administration
and provides an overview of the international landscape of generative AI. In Finnish with abstract in English. - HUDERIA MethodologyOpens in a new window. (Alan Turing Institute, 5.12.2024)
A multi-step governance process developed by the Alan Turing Institute and adopted by the Council of Europe. It provides an evidence-based and structured approach to carrying out risk and impact assessments for AI systems. Link to the method document is in the article. - Demystifying Local and Global Fairness Trade-offs in Federated Learning Using Partial Information DecompositionOpens in a new window. (Faisal Hamman, Uni. of Maryland, 7.9.2024)
"As machine learning permeates high-stakes sectors like finance, healthcare, recommendation systems, etc., ensuring these algorithms are fair becomes increasingly crucial. This issue is especially complex in the federated learning setting." This summary article contains the link to the original research paper. - Why bigger is not always better in AI Opens in a new window.(MIT Technology Review, 1.10.2024).
An article on the development and proven benefits of smaller AI models compared to very large models. - Tekoälyn vastuullisuuskeskustelussa kaivataan sidosryhmien ääntäOpens in a new window. (Marika Lanne, VTT. Published on the ETAIROS research project blog).
The article explores the benefits and methods of active stakeholder engagement, which is important in AI development. In Finnish. - Generative AI Service Demo to Support Law Drafting of the Ministry of Transport and Communications, final reportOpens in a new window. (Publications of the Ministry of Transport and Communications 2024:9).
A comprehensive, detailed report on the experiment to assess the suitability of Finnish large language models as a tool for drafting legislation. Also includes basic information on the processes and concepts associated with large language models. The raport is in Finnish but includes a summary in English. - What is retrieval-augmented generation (RAG) and what does it do for generative AI? Opens in a new window.(Nicole Choi, GitHub, 4.4.2024).
A practical article on the RAG method to facilitate the useful use of large language models. - Are Emergent Abilities in Large Language Models just In-Context Learning? (Sheng, Bigouleava et al., Tech. Uni. of Darmstadt, Uni. of Bath, August 2024).Opens in a new window.
The study investigated the extent to which large language models have emergent properties in the autonomous understanding of natural language. - What are the risks from Artificial Intelligence?Opens in a new window. A comprehensive living database of over 700 AI risks categorized by their cause and risk domain. (Massachusetts Institute of Technology MIT, 14.8.2024)Opens in a new window.
- Tekoäly tuli jo työpaikoille (Panu Räty, TiVi, 18.8.2024)Opens in a new window..
News article about the internal use of large language models in Finnish companies. In Finnish, for TiVi subscribers. - Tekoälyn nopea kehitys sai myös yritysjätit huolestumaan – parissa vuodessa on tapahtunut iso muutos (Mikko Asunta, YLE, 18.8.2024)Opens in a new window..
"As many as 56% of the largest 500 companies in the US cited AI as a risk factor in their annual report last year. In 2022, the figure was only 9%." The story is in Finnish only. - Botti, rakastettuniOpens in a new window. (Albert Koski, Helsingin Sanomat, 5.8.2024).
Wide-ranging feature explores the dilemma of AI friends and partners. In Finnish, for subscribers of Helsingin Sanomat. - We need to prepare for addictive intelligenceOpens in a new window. (Robert Mahari & Pat Pataranutaporn / MIT Technology Review, 5.8.2024).
The article analyses the human tendency to project human characteristics onto devices, currently in particular AI applications that mimic human behaviour and conversation. - Unmasking Secret Cyborgs and Illuminating Their AI ShadowsOpens in a new window. (Abhishek Gupta, Boston Consulting Group / TechPolicy 7.6.2024).
Ethical challenges and practical risks of integrated and even hidden AI elements in widely used information systems. - Tekoäly tuli jo työnhakuun – ”oman osaamisen kuvaaminen voi olla vaikeaa”Opens in a new window. (TiVi 29.7.2024).
News story on the development, use and challenges of AI in the TE-Services' Job Market Finland web service. (Subscribers and in Finnish only. Check if your organisation offers access to Alma Media's publications.) - Best Practices in AI Documentation: The Imperative of Evidence from PracticeOpens in a new window. (Amy Winecoff & Miranda Bogen / Center for Democracy & Technology, 07/2024).
Benefits and methods for accurate documentation of artificial intelligence systems. - AI trained on AI garbage spits out AI garbageOpens in a new window. (Scott J Mulligan / MIT Technology Review, 07/2024).
What happens when AI is trained on AI-generated data? The latest research on the subject. - AI in the public sector: white heat or hot air?Opens in a new window. (Imogen Parker & Matt Davies / Ada Lovelace Institute, 07/2024).
An overview of the opportunities and challenges of AI in the public sector in the UK, largely relevant to the Finnish situation as well. - AI can make you more creative—but it has limitsOpens in a new window. (Rhiannon Williams / MIT Technology Review, 07/2024).
News article about a study on the impact of generative AI on human creativity. - Exposing The Brittleness Of Generative AIOpens in a new window. (Lance B. Eliot / Forbes, 02/2024).
Analysis of error-proneness in generative AI in the light of a failure event in ChatGPT. For subscribers, but possibly available as a free sample article. - Meaningful Public Participation and AIOpens in a new window. (Anna Colom / Ada Lovelace Institute, 02/2024).
A concise overview of the importance of citizen participation in AI development. - Scientists' Perspectives on the Potential for Generative AI in their FieldsOpens in a new window. (Meredith Ringel Morris / Google Research, 04/2023).
An overview of the potential of generative AI in different domains, based on extensive researcher interviews.
Updated: 4/3/2026