AI Policy Lab Day 2025: Highlights and Reflections (Recording Available)
Date & Location: November 19, 2025, Umeå University, Västerbotten, Sweden
The AI Policy Lab Day 2025 was rich in insight and exchange.
Sennay Ghebreab delivered a keynote that grounded Question Zero in lived experience, reminding us that the decision to use, or not use AI is never a static checkpoint. He urged us to think in terms of Question Infinity:a continuous, reflective process in which risks and opportunities are held in tension rather than framed as opposites.
Daniel McQuillan‘s talk added a powerful systemic lens. By framing contemporary AI as a product of deeper structural failures, he challenged us to confront the material and social realities beneath technological optimism. His proposal of decomputing, a combination of degrowth, conviviality, and care, called us to imagine responses that prioritise collective well-being over speed or scale.
Our researchers’ posters reflected a striking level of maturity. Their work is rigorous, thoughtful, and already influencing wider debates on responsible AI. It was encouraging to see how confidently they engaged with participants and how deeply their projects connected to real societal needs (Rachele Carli, Petter Ericson, Jason Tucker, Tatjana Titareva, Themis-Dimitra Xanthopoulou, PhD Mattias Brännström).
Throughout the afternoon, participants brought curiosity, openness, and an eagerness to engage in discussions and informal exchanges between sessions.
The evening screening of Humans in the Loop added an emotional and narrative dimension that tied the day together. The dramatized story, rooted in the real experiences of data workers in India, wove together the daily realities of annotation labour with local culture, personal aspiration, and the power of lived experience. It captured the invisibility of this global workforce while honouring their agency and resilience. The discussion that followed made clear how crucial these perspectives are for any serious conversation on responsible AI.
A full day of insight, critical dialogue, and shared commitment.
Responsible AI Self-assessment Workshop: Start with Question Zero
Date & Location: August 27, 2025, Umeå University, Västerbotten, Sweden
On 27 August 2025, more than 100 participants joined the AI Policy Lab workshop Responsible AI Self-Assessment: Start with Question Zero at Umeå University and online. Together, we tested and debated the Responsible AI Self-Assessment Tool, designed to help organisations pause, reflect, and ask why before moving into AI adoption.
The workshop brought together voices from academia, industry and the public sector, sparking vibrant conversations around responsible AI. Participants reflected on questions such as:
Should a clear AI clarification step be required before entering Question Zero (“Why do you plan to adopt an AI system?“)
Should organisations complete a process pre-assessment before starting with AI?
What kind of work should remain human-only?
How can transparency and ethics be maintained when deciding between automation and augmentation?
Why might other non-AI solutions not solve the problem at hand?
We are deeply grateful to everyone who joined, shared perspectives and challenged assumptions. Your input is vital to shaping a practical, responsible approach to AI adoption.
Next steps
The tool is still a work in progress. Feedback from this workshop will be implemented directly into the next version of the tool. Future workshops will continue to stress-test and evolve it, ensuring it meets the needs of diverse organisations working with AI.
As Virginia Dignum, Director of the AI Policy Lab, put it:
“Responsible AI isn’t AI-first, it’s people-first. It starts by asking why, not rushing to deploy.”
Interested in taking part in upcoming sessions? Keep an eye on our website and LinkedIn page for updates.
Global AI Policy Research Network Launched at UN IGF 2025 (Recording available)
Workshop #288: An AI Policy Research Roadmap for Evidence-Based AI Policy Date & Location: June 26, 2025, Oslo, Norway
At the UN’s Internet Governance Forum (IGF) 2025 in Oslo, Norway, AI Policy Lab @Umeå University (Virginia Dignum, Jason Tucker, Tatjana Titareva and colleagues) and Mila – Quebec Artificial Intelligence Institute (Isadora Hellegren Létourneau, and colleagues), in cooperation with our partners including Alex Moltzau, Eltjo Poort, Neema K. Lugangira, and many others, launched the Global AI Policy Research Network (GlobAIPol). The network invites diverse stakeholders to share practical knowledge that supports ethical, transparent, and evidence-based practices for shaping inclusive and trustworthy AI policies. The session also encouraged global stakeholders to endorse the Roadmap for AI Policy Research.
“AI does not happen to us! AI is designed by humans. We make the choices.” – Professor Virginia Dignum’s keynote reminded us that before asking how to implement AI, we must ask Question Zero: Is AI the best option here? We need to shift from fragmented, reactive policies to coordinated, evidence-based strategies rooted in ethics and justice.
The interventions and discussion revealed critical lessons from global perspectives:
The EU is demonstrating promising approaches with the European AI Office expanding from 97 to 140 staff by the end of 2025, supporting regulatory sandboxes and international collaboration including a €5 million generative AI initiative with Africa.
In healthcare, we must move beyond treating AI as a “magic pill” and build upon existing regulatory frameworks – just as we trust paracetamol today because of rigorous oversight developed several decades ago.
Well-designed regulation stimulates innovation rather than slows it down. Different countries need diverse legislative approaches harmonised with local values, not a one-size-fits-all global AI governance structure.
The time to act is now. AI is shaping our collective future, and how we act today will define who benefits, who is heard, and who is left behind.
AI Technologies in Public Service: A Workshop for Identifying Needs, Challenges, and Solutions
Date & Location: April 9, 2025, Umeå University, Västerbotten, Sweden
The workshop was organised by the AI Policy Lab in collaboration with the AI Technologies for Sustainable Public Service Co-creation (AICOSERV) project members.
Overview
The workshop brought together more than 50 stakeholders from the public and private sectors, as well as academia, to explore the relationship between barriers to AI adoption in public services and the skills and expertise required to overcome them.
A central theme of the workshop was the “question zero”, the fundamental inquiry of whether AI should be used at all in a given context. As AI technologies continue to advance and expand into complex public sector tasks, the assumption that AI is always the right or necessary solution must be critically examined. The workshop challenged participants to consider not only how AI can be implemented, but to question whether it should be, emphasizing that responsible adoption begins with questioning the appropriateness and desirability of AI in specific domains.
This foundational concern set the tone for broader discussions about trust, governance, transparency, and the skillsets required to navigate the opportunities and risks of AI in public service.
Keynote Address
Professor Virginia Dignum, Director of the AI Policy Lab, opened the workshop with a keynote titled “Governing AI: Why, What, How?”
She addressed the societal and governance implications of AI, focusing on the need to critically evaluate when and how AI should be integrated into public service contexts. Her talk stressed the importance of not overlooking ethical, legal, and operational challenges in the rush to adopt AI.
Regional Case Study: AI in Västerbotten
Considering the wide range of public services open to AI adoption, a recurring set of challenges consistently emerges. Whether deploying AI-driven diagnostic tools in healthcare or implementing predictive analytics within smart city infrastructures, public and private sector actors, and community stakeholders face diverse barriers. Henry Lopez-Vega, fellow at the AI Policy Lab, presented on the challenges of AI adoption in the Västerbotten region in his session titled “What are the challenges with AI (in Västerbotten)?”
His research identified three core barriers to building a responsible AI ecosystem:
Technological infrastructure and processes within organisations
Organisational culture and resistance to change
Lack of clarity around AI governance and ownership
Group Discussions: Skills and Stakeholder Engagement
In the second half of the workshop, participants engaged in group discussions focusing on organisational challenges, key stakeholders, and barriers to implementation. Each group then mapped the skills and knowledge needed for responsible AI adoption in their contexts.
For example, in the case of AI-supported recruitment processes, participants identified several critical barriers:
Lack of transparency in AI decision-making
Biases in training data
Limited legal and ethical guidelines for automated hiring
To address these issues, participants emphasized the need for professionals with a blend of competences, including:
Knowledge of data protection and anti-discrimination legislation
Skills in evaluating and auditing AI systems
Awareness of ethical considerations in algorithmic decision-making
Findings and Framework
The increasing efforts to implement AI across various public sector domains have led to a critical question: what types of professionals, equipped with what specific skills and competences, should lead the integration of responsible AI? Defining the essential set of skills, knowledge, and professional competences required for the effective and ethical deployment of AI in both public and private sector services becomes a key priority.
A key outcome of the workshop was the identification of a recurring set of challenges affecting. These include:
Low levels of trust
Lack of transparency
Unclear ownership and responsibility
Insufficient stakeholder awareness
Limited AI literacy and governance skills
To address these challenges, we propose the conceptual framework depicted in Figure 1. This framework highlights the urgent need to cultivate professionals who combine technical expertise, ethical sensitivity, and domain-specific knowledge to lead responsible AI integration. It maps the interconnected layers that influence the deployment and responsible use of AI technologies in public services, including:
Contextual Challenges – such as low trust, resistance to change, and limited organisational readiness
Structural Barriers – including unclear project ownership, inadequate governance frameworks, and insufficient infrastructure
Skill and Knowledge Gaps – highlighting the lack of AI literacy, ethical awareness, and domain-specific competences
Stakeholder Roles – outlining the importance of identifying and engaging relevant actors (e.g. policymakers, IT professionals, legal advisors, and citizens) throughout the AI lifecycle
This framework is intended to guide structured reflection and planning around AI deployment, helping ensure that technologies are not only functional but also trustworthy, inclusive, and aligned with public values. As such, it can serve as a practical tool for organisations seeking to integrate AI technologies responsibly. It encourages a systemic perspective, one that moves beyond technical feasibility to consider broader organisational, social, and ethical dimensions.
By applying this framework, decision-makers and project leads can:
Identify context-specific challenges before adopting AI
Map key stakeholders and clarify roles and responsibilities
Recognise skill and competence gaps that must be addressed
Design AI initiatives that align with principles of transparency, accountability, and fairness
Figure 1. The framework summarising the workshop’s thematic discussions
In conclusion, the workshop underscored that a stakeholder-oriented, challenge-driven approach is key to enabling responsible AI adoption. By starting with specific domain needs and mapping corresponding skills and knowledge, organisations can more effectively navigate the complex landscape of AI integration.
Participants discussed responsible AI from multiple perspectives – technological, ethical, and social. Below we share the key messages from the retreat. We hope to inspire future multidisciplinary discussions, workshops and projects related to responsible AI research, literacy, and practical solutions for diverse stakeholders in Sweden and internationally.
Question Zero & Human Responsibility
When we hear about new AI ventures, we need to ask the question zero: “Is the adoption of an AI tool the best solution for our current problem?” with assessments including the environmental costs of running this AI system and understanding of what specific aspects of operations it improves. This includes critically assessing who is participating, what is the provenance and the management of data, what are the bases for modelling and design choices, how will results and impact be evaluated. AI does not happen to us, we [humans] design and/or adopt and use the AI systems.
Responsible AI development should follow a compositional approach, where verified datasets and models with clear principles can be combined to create new systems. This framework emphasises the need to balance ethical considerations and accuracy while accounting for differences across sectors, industries, and global value systems. The approach prioritises sustainability, transparency, control, regulation, participation, inclusiveness, trust, and fairness, with a focus on measuring societal implicationsand ensuring equal representation and identification.
True AI fairness goes beyond mathematical equality, considering historical disparities and contextual factors, and diverse definitions of fairness, as mathematical fairness alone can still create discriminatory outcomes in complex social settings.
Participation, where people meaningfully contribute to (design) decisions, strengthens democracy and supports responsible design, development, and use of AI. However, participation alone does not guarantee better outcomes; thoughtful design is needed to prevent manipulation and address blind spots by incorporating diverse perspectives.
Technological change is ecological, not additive, in that it is transformative and systemic rather than simply incremental, fundamentally reshaping existing environments and society. Drawing from Neil Postman’s (1995) framework, technological innovations fundamentally transform existing systems rather than simply adding new capabilities, creating trade-offs, winners and losers, and often leading to reframing how we perceive the world.
Structural challenges. Several “traps” hinder responsible AI research: the framing trap (failure to model entire systems), portability trap (ignoring context sensitivity), formalism trap (oversimplifying social concepts), ripple effect trap (missing ecological impacts), and solutionism trap (overreliance on technological fixes).
AI Art. The concept of “AI as a mirror” raises important questions about artistic expression, with significant differences between the inner experience of human artistic creation versus AI-assisted art generation, challenging traditional perspectives on IP rights, creativity (human only, genAI only, human-AI hybrid), intention, and artistic value.
If the ideas above resonate with you, we encourage to check out the AI Policy Lab LinkedIn page for common cooperation opportunities.
We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.
