The meetings will take place on the 21th October 2026, 9:00 to 16:00 at Domus Juridica, Kristian Augusts gate 17, Oslo.
.png)
Research-areas that are organising a meeting (all meetings are open to everyone in TRUST who is interested in the specific area):
Causal AI (research area 4)
Causal AI tries to understand cause and effect, not just patterns. This area develops methods that let AI reason more like humans — identifying how one action leads to another. This makes AI fairer, more transparent and more reliable in sensitive areas such as law, medicine and policy.
Meetingorganiser: Lana Bubalo
Representation and alignmen (research area 5)
The meetings will have presentations of projects connected to this RA. Projects will be briefly introduced, focusing on how it connects to alignment and/or representation, and mention any relevant outcomes, questions, or aspects that could be useful for the wider RA5 discussion. After the short presentations, we will have the possibility to discuss how the different projects connect, what is already covered, what may be missing, andpossible areasfor future research and collaboration.
Meetingorganiser: Guendalina Righetti
Foundation models (research area 6)
Large general-purpose models, such as language models, can be adapted for many uses. TRUST focuses on making these models more efficient, sustainable and ethically sound. The research looks at smaller and greener versions, new training methods and clear rules for responsible fine-tuning.
Meeting organizer: Filippo Bianchi
Generative AI (research area 7)
Generative AI creates new content such as text, images or sound. This area studies how such systems can be designed and used responsibly. TRUST explores ways to make AI-generated material identifiable, prevent misinformation and encourage creative cooperation between people and machines.
Meeting organizer: Pierre Lison
Emergent behaviours including multi-agents (research area 8)
[text coming]
Emergent behaviour and cognition (research area 8)
Background: When powerful new AI tools and infrastructures are designed and used, we need to understand and explain how cognition emerges and how this influences the division of labor between humans and AI (including different types of automation and augmentation) in multiple types of settings. As AI systems are used and scaled, they also become cultural tools that affect both micro- and macro-level dynamics. AI can create different types of disruptions in cognitive and social processes.
New conceptualizations are needed to capture the interaction between AI and humans. The concept and stance complementary intelligence offers a position from cognitive AI and machine AI perspectives and seems promising. What’s an adequate unit of analysis should also be re-conceptualized.
In this seminar/workshop, we will present and discuss a number of projects-pre-projects that address research questions related to memorability and emerging cognition and learning in multiple types of settings.
RA 8 is also related to other RA and several AC.
Presenters: Thomas Espeseth, PSI, UiO; Stefan Schauber, Faculty of Medicine, UiO,Cato Bjørkli PSI, UiO; Crina Damsa IPED, UiO; Sten Ludvigsen, IPED UiO
Invited participants: Omid Mirmotahari and Yngvar Berg IFI, UiO; Andreas Brandsæter Volda University college, Anders Malthe-Sørensen, Dept- of Physics UiO, Malcolm Langford, Faculty of Law and TRUST leader.
We will involve also other participants across field of knowledge and institutions based on interest.
Programme:
9.00-9.05: Introduction; Sten Ludvigsen
9.05-9.30: Attribution, Memorability, and Understanding in Human-AI Interaction: Thomas Espeseth. PSI, UiO
9.30-9.55: An Evidence-Based Model to Boost Learning in Higher Education (EMBLA): Stefan Schubert. Medical faculty, UiO.
Break
10.05-10.20: SimAI: AI-enhanced Learning of Team Skills in Education for High-risk Professions. Crina Damsa. IPED, UiO.
10.20-10.35: Emerging cognition in multiple settings: Cato Bjørkli. PSI, UiO.
Break
10.35-10.55: General discussion and discussion about “The Idea about a Position paper”. Individual and collective cognition: Human AI and Machine AI. Sten Ludvigsen. IPED, UiO.
10.55-11.00: Next steps
Meeting organiser: Sten Ludvigsen
Disruptive AI-driver digitalisation (research area 12)
The meeting will bring together TRUST researchers interested in RA12 to share ongoing and planned projects related to Disruptive AI-DrivenDigitalisation.
Meeting organiser: Mike Barnes
Reforming AI governance (research area 13)
This is an open meeting to discuss current AI governancein light ofexisting or future TRUST center projects, current events and developments, and partner needs. If you wish to give a 3 min lightning talk, feel free to register here.
The meeting willat the same time be a meeting of Action Cluster 8 - National and international governance for trustworthy AI.
Meeting organiser: Tobias Mahler
Action cluster meetings (all are open to everyone in TRUST with an interest in the topic).
AC6: Combatting criminality with AI
Combatting criminality with AI seeks to enable data aggregation, behaviour prediction, and selected real-time AI responses to AI-driven criminality, while addressing concerns about privacy, explainability, security, and fairness (e.g., predictive policing and discrimination). We will test: (i) temporal-spatial foundation and anomaly detection models on financial and surveillance data to create digital evidence; (ii) the efficacy and sociological effects of responsible UQ and causal explainability in policing settings; and (iii) how to balance privacy and discrimination for high-risk AI in practice.
Meeting organiser: Helene Gundhus
AC13: Green AI for green transitions
A joint status check on submitted user cases and the efforts to include them in research projects. Discussion of sustainability-related reseach project (e.g., AI-ST) and their relevance to user partners.
Meetingorganiser:Tuukka Mäkitie
Meetings arranged by projects (click the title to read more about the project):
Most projects meeting are open to everybody, while a few are working sessions on invitation only.
Project 12 - TruSTFM - Foundations for a Trustworthy Spatio-Temporal Model (by invitation only)
Meeting organiser: Filippo Bianchi
Project 15 - Trustworthy AI for the Peace and Security Domain
We invite to an open meeting to explore potential collaborations on TrustworthyAI4Peace & Security.
About the project: AI is increasingly utilized to support international security, conflict prevention, and peace building, where mistrust is a widespread challenge – for instance, among conflict parties and between conflict parties and mediators or peacekeepers. Despite this, little is known about what makes AI applications employed in this area more trustworthy. There is nocommon understandingof which principles should guide AI use and how to balance priorities such as accuracy, transparency, inclusivity, and explainability. This project aims to fill these gaps by developing a comprehensive model of TrustedAI4Peace&Security, along with practical criteria and measures for responsible AI design and deployment.
About the meeting: Goingforward, we plan to collaborate directly with user partners and AI researchers to explore specific use cases. Do you want to know more about the project? Do you know about potential use cases and AI technologies that we could explore with our international partners? Or are you an expert in AI ethics, AI regulation, or another domain that could be relevant for our project? If so, please join us! We will provide an overview of current AI uses in peacebuilding and conflict prevention and thenfacilitatea brainstorming session.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser:Andreas Hirblinger
Project 31 - Prototype to Product: Integrating AI into Complex Systems (P2P-AI)
Prototype to Product: Integrating AI into Complex Systems (P2P AI) focuses on how to successfully move AI from small scale prototypes into everyday use in large organizations and complex IT environments. The project develops practical guidelines and an agile process for integrating AI modules step by step, helping teams balance key factors like maintainability, accuracy, scalability, security, and user experience, and showing when to involve legal, domain, and ethics experts. Starting with real cases at the HISP Center atUiOand extending to TRUST projects and Action Clusters, P2P AI looks for concrete success factors and lessons learned, so users and organizations can reduce risk, avoid common pitfalls, and achieve trustworthy, effective AI integrations in practice.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser: Antonio Martini
Project 33 - Digital sovereignty and open-source LLMs in the public sector
Public sector organizations are adopting large language models, typically defaulting to commercial platform providers; decisions that risk locking in vendors and governance frameworks. This project investigates how open-source LLMs can become a genuine alternative, offering transparency, auditability, and digital sovereignty. We examine the pathway to viability across four dimensions - technical, organizational, policy, and international politics. The research is urgent: as implementation decisions accumulate switching costs, the window forestablishingviablealternatives is closing.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser: Alexander Kempton
Project 34 - DT4GeoRisk (by invitation only)
Meeting organiser: Zhongqiang Liu
Project 48 - Case processing with AI (CASPAI)
CASPAI (Case Processing with AI) will develop and test an interdisciplinary meta-framework for trustworthy AI in public decision-making, using asylum adjudication as a deliberate "hard case". Asylum falls within four of the EU AI Act's high-risk categories and concerns highly vulnerable people, so solutions that work there are designed togeneraliseacross public administration and the courts, including TRUST user partners. The projectbuilds ona research infrastructure of almost one million Nordic-Baltic asylumdecisions, andruns alongside theNordForsk-fundedNordAIdproject (Oslo, Copenhagen, Uppsala, Vilnius). Three questions guide it: when can AI-based public decision-making be trustworthy, how can trustworthiness be enhanced, and do these innovations produce calibrated trust – neitherovertrustnorundertrust? Five work packages cover veridical and privacy-secure AI; anticipatory governance under the AI Act; a spectrum of decision-support models, from legal source recommenders to generative drafting and agentic AI, plus a preparation tool for asylum seekers; explainable AI and uncertainty quantification; and responsible evidence-based innovation through two-phase experiments with stakeholders in four countries.Participatory design and shared value metrics run throughout.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser: Malcolm Langford
Project 51: PROMESS (Processing the Mess: Critical, Ethical & Legal Exploration of AI Tools in Policing)
This is a transdisciplinary project responding to today’s overload of messy information in police investigations and intelligence, where information management remains largely manual and time-consuming and computational methods are still limited. Using real police data, the project develops and evaluates machine learning tools to transform unstructured information into structured, searchable knowledge, including graph databases and tools to detect conflicting or missing information. In parallel, PROMESS traces how these technologies are domesticated and co-produced in everyday policing, examines how they affect professional judgement and critical thinking, and assesses their compliance with European legal frameworks, human rightsstandardsand emerging AI regulations. Working closely with user partners such asKripos, the project aims to bridge the gap between current investigative and intelligence practices and modern computational processing, delivering both practical applications for more efficient and trustworthy information handling, and clear guidelines for the responsible, transparent and accountable use of AI in policing.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser: Helene Gundhus
Project 60 - School Mathematics in the Age of AI
This meeting aims for aninteractive discussion on the future of school mathematics in an AI-rich society, aimed at shaping a new PhD project (currently in the recruitment phase) and building collaborations across TRUST.If AI can solve equations, write code, analyze data, and even perform mathematical modelling, what is left for students to learn? This workshop invites participants to discuss the future role of school mathematics and the competencies needed in an AI-rich society.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser: Andreas Brandsæter
Project 72 - Generic trend detection (by invitation)
Meeting organiser: Johannes Voll Kolstø
Project 80 - Agentic Sim-Based Inference under Computational and Data-Quality Constraints: Digital Twins for Healthcare
We start a project on Agentic AI in situations where there are digital twins or simulators - often based on both domain knowledge and machine learning components - of different quality, precision, and computational complexity, and where there are multiple types of data. If you have similar problems and interests, please join our meeting, to makea contact.
This meeting is open to everyone in TRUST with an interest in the topic.
Meeting organiser: Ole Jakob Elle