Policy briefs: New Challenges to Democracy

Far-right populism and decarbonization in Europe 

By Mahir Yazar and Håvard Haarstad, University of Bergen 

Download policy brief in pdf

Summary/background

Carbon-intensive industries are concentrated in a few carbon-intensive regions in Europe. Fossil Fuel phaseout, referred to as decarbonization, actions will affect those regions particularly strongly. Correspondingly, carbon-intensive regions often exert significant political influence on decarbonization policies and actions both at the national and regional levels. Populist far-right parties remain potent players delaying rapid energy transition in the carbon-intensive regions of Europe.

This policy brief focuses on three countries, namely Estonia, Germany and Poland, and the populist far-right parties´ discursive-institutional influence on the regional decarbonization strategies. Scientists at CET present evidence of how populist far-right parties in these countries politicize decarbonization by infusing disinformation and attacking institutions that could accelerate regional decarbonization. This policy brief emphasizes the importance of understanding populist far-right discursive tactics to prevent politicization of decarbonization and climate change at the regional level and ensure effective policy implementation of national and international climate policies.

Key findings

The Conservative People’s Party of Estonia (EKRE), the Alternative for Germany (AfD), and Law and Justice (PiS) in Poland have constructed regional decarbonization as a direct threat to national sovereignty, and the elements of regional identity, such as unique family and cultural traditions. These themes increase tensions between regional and national policies and undermine efforts to create a unified approach towards decarbonization in the carbon-intensive regions of Ida-Virumaa, North Rhine-Westphalia, and Upper Silesia.

The populist backlashes against the EU-driven progressive public policies and anti-democratic rhetoric, including xenophobia and national sovereignty discourses are commonly used by these three populist far-right parties to mobilize counternarratives against climate change and regional decarbonization. EKRE and PiS typically portray themselves as the protectors of family values, social insurance and safety for
mine workers affected by regional decarbonization. PiS and AfD harness regional identity to mobilize civic engagement against decarbonization. All three parties work to dismantle key decarbonization institutions, either by holding key ministries to delay policies or by collaborating with climatedenial organizations for disinformation targeted at key national institutions.

Scientists at the Centre for Climate and Energy Transformation at the University of Bergen identify three discursive-institutional tactics used by populist far-right actors to delay decarbonization: (1) politicizing decarbonization, (2) reframing cultural values to form alliances with anti-decarbonization movements, and (3) dismantling key decarbonization institutions.

Photo:

CET

Overall, carbon-intensive regions in Europe are particularly susceptible to the discursive tactics and institutional work of populist far-right parties, and may therefore provide opportunities for these parties to constrain decarbonization more broadly.

Recommendations for policy-makers

• Policymakers must develop clear communication strategies emphasizing the longer-term benefits of decarbonization on improved environmental and health conditions and creating new labour skills in the increasing job market for clean-energy sectors.
   
• To counter disinformation, policymakers must ensure transparency and public participation in the carbon-intensive regions and share energy transformation planning processes with the public to increase trust and visibility of the potential long-term benefits of the energy transformation.
   
• Significant investments in public education campaigns are essential to raising awareness about climate change, renewable energy, sustainability, public health, and the importance of learning new skills for the highly competitive European job market.
   
• The legal and institutional frameworks must be strengthened to support regional decarbonizationefforts. EU funding agencies must navigate ways to directly fund and support regional institutions promoting decarbonization, such as research centers, innovation hubs, or training programs for green jobs, to ease the burden of transition.
   
• International cooperation and knowledge sharing must be fostered among regions leading the way in decarbonization efforts to create a sense of solidarity and shared responsibility in creating a sustainable, prosperous future for all.

Authors

Mahir Yazar, Centre for Climate and Energy Transformation (CET), University of Bergen
Håvard Haarstad, Centre for Climate and Energy Transformation (CET), University of Bergen

Related article

Populist far right discursive-institutional tactics in European regional decarbonization.
Political Geography,105, 102936. https://doi.org/10.1016/j.polgeo.2023.102936

Do consultations matter for stakeholder support for policy proposals? Evidence from the European Union

By Adriana Bunea, University of Bergen

Download policy brief in pdf

Summary and recommendations

A key assumption underpinning participatory governance is that there is a positive relationship between stakeholder participation in policymaking and the levels of public legitimacy and stakeholder support for policies formulated with the help of stakeholder engagement. However, this fundamental assumption is only rarely tested empirically with the help of data coming from actual stakeholder evaluations of concrete policy proposals. 

We address this gap and ask: do consultations matter for stakeholders’ support for policies formulated through stakeholder participation in EU policymaking? 

We find that proposals formulated with the help of open public consultations generate systematically higher levels of stakeholder support. Providing more information about the consultation activities reduces stakeholder support. There is no systematic association between policymakers’ attention to stakeholders’ consultation inputs and stakeholders’ support for proposals.  

Recommendations

• Organising open public consultations and informing the public about them may enhance the process and outcome legitimacy of policy proposals. 
   
• Policymakers should carefully consider the amount and type of information they publicly communicate about consultations and convey this information in an easily accessible format
   
• Policymakers should actively work towards understanding what stakeholders consider to be a ‘procedurally fair’ consultations, and design them accordingly to enhance the process and output legitimacy of their decision-making  

Stakeholder consultations

Stakeholder participation in the formulation of public policies is a landmark of modern governance. For non-elected institutions, such as national and international bureaucracies and regulatory agencies, stakeholder engagement is particularly relevant. 

Bureaucracies use public participation to gather information and to build support for proposed policies. This aims to ensure the successful implementation of policies and to strengthen bureaucratic institutional legitimacy, autonomy, and power. A key assumption is that there is a positive relationship between stakeholder participation and the levels of public legitimacy and stakeholder support for policies formulated with the help of stakeholder engagement.  

However, this assumption is rarely tested empirically. The research examining this relationship employs mainly survey experiments that may lack external validity and do not account for policy complexities. Systematic analyses of how consultations shape policy support in observational studies that use actual stakeholder evaluations of concrete policy proposals formulated with the help of different consultation activities, across policy areas, are currently missing. 

Addressing this research gap is relevant for several reasons. Stakeholder participation in policymaking is crucial for better regulation policies. Understanding how stakeholder engagement contributes towards formulating policies that enjoy stakeholder support is key to understanding how better regulation policies generate policy processes and outcomes that enjoy public legitimacy. The impact of consultations on stakeholder support is vital for debates on democratic innovations, transparency, and stakeholder inclusiveness in bureaucratic policymaking. In EU policymaking, the information about stakeholders’ support for draft policy proposals feeds directly into the legislative decision-making process. Understanding the conditions under which EC legislative proposals generate stakeholder support is relevant for understanding the EU legislative decision-making process. 

Our research 

We analyse a new dataset assembled as part of the ERC Starting grant project CONSULTATIONEFFECTS. It contains information about the consultation activities employed by the European Commission to formulate 316 policy proposals during 2016-2021, across all policy areas, on which 8,955 stakeholder evaluations were received. We examine the correlation between indicators describing how policymakers describe consultation activities in the ‘Explanatory memorandum’ preceding the policy proposal and stakeholders’ support for the proposal.

Building on procedural fairness theory, we hypothesize that proposals on which policymakers communicate that were formulated through an inclusive consultation process, on which they provide more information about consultation activities and on which they show more attention to stakeholder consultation inputs, garner higher levels of support. 

We employ sentiment analysis of feedback texts submitted by European stakeholders on the EU webpage ‘Have your say!” to measure stakeholder support. The variables measuring consultation inclusiveness, transparency, and attention to stakeholder inputs were coded based on the ‘Explanatory memorandum’ of each legislative proposal. We use regression analysis to examine the relationship between EC policymakers’ efforts to communicate information about the procedural fairness of consultations and levels of stakeholders’ support. 

Figure 1.

Photo:

Adriana Bunea

Figure 1 illustrates the results of our analysis. It shows that proposals mentioning an open public consultation during their formulation stage are significantly more likely to receive more positive stakeholders’ evaluations. However, increased transparency about consultation activities (more information) leads to lower stakeholder support. There is no systematic association between policymakers’ attentiveness to consultation inputs and levels of support. These findings highlight the legitimizing power of open consultations and the mixed effects of transparency on process and output legitimacy

Policy implications

Our findings suggest that the European Commission’s efforts to expand stakeholders’ participation in EU policymaking through continuous reforms of the better regulation policy may have a lower impact on enhancing legitimacy than initially hoped for. European policymakers should invest time and resources in organizing open public consultation and inform the public about them. The current approach employed by the Commission to communicate about consultations during the post-decision stage works suboptimal in terms of convincing external audiences about the fairness of the consultations.

Policymakers should carefully consider the amount and type of information they provide publicly about consultations and convey this in accessible formulation and format. Our findings suggest a disjoint between what policymakers consider important to justify their decisions and what stakeholders consider important to justify and accept these decisions. Policymakers should better understand what stakeholders consider to be a ‘procedurally fair’ consultation process and design them accordingly. This could enhance both process and outcome legitimacy.  

References

Full details of the study mentioned in the policy brief:
Adriana Bunea and Idunn Nørbech (2023) ‘Do consultations matter for stakeholder support for policy proposals? Evidence from the European Union.’ Working paper.
ERC Starting Grant CONSULTATIONEFFECTS website

Author

Professor Adriana Bunea, PI of ERC Starting Grant CONSULTATIONEFFECTS, Department of Comparative Politics, UiB

Climate emotions – and how they can motivate or hinder action

By Thea Gregersen, NORCE

Download policy brief in pdf

Summary 

People’s emotions related to climate change and climate solutions are highly relevant for their climate attitudes and engagement, and for how they perceive and interpret new climate change information. Emotions such as guilt, anger, hope, fear, and sadness can motivate or hinder action - unite or polarize. Of special interest to democracy is the potential polarizing or de-polarizing effect of emotions, as well as the role of emotions in voting decisions, policy support and people’s trust in democratic processes. Studies on climate emotions can help us understand and predict opposition and conflict and contribute to tailored communication campaigns.

The issue 

To meet the climate targets and adapt to a changing environment, democratic societies depend on citizens’ willingness to accept new policies and technologies and (for some groups) change their behaviors. People’s emotions related to climate change can help or hinder mitigation efforts. For example, worrying about the wellbeing of oneself and others can motivate willingness to reduce one’s carbon footprint and climate anger can mobilize movements, but might also fuel reactance and polarization.

Emotions can reveal how people perceive a situation and help us predict how they will act. People’s emotional reactions depend on their appraisal of a situation (e.g., sadness in response to experiencing irrevocable loss, fear in response to an uncontrollable threat) and motivate behavioral reactions. Different emotions are related to different action tendencies and most emotions can lead to more or less constructive reactions (e.g., sadness can lead to withdrawal or to helping behaviors, fear to fight, flight, or freeze). People are not likely to experience only one isolated emotion in relation to a situation or phenomenon, further complicating the links to cognition and motivation. 

Emotions also influence how information is evaluated. Emotion-inducing communication strategies (e.g., tailoring messages aimed at eliciting fear or hope) and displays of emotions by activists, researchers or citizens can influence the emotions, thoughts, and behaviors of the audience. However, the same information, situation, or emotion display can give rise to hope in some people and anger in others, depending on their existing emotions, values, and beliefs. In sum, the motivating potential of emotions is complex.

Our research 

• Studies looking into specific causes of climate emotions have found that their effects are not straight forward. Emotions such as hope and anger might relate to both higher and lower climate change engagement – depending on who or what the emotion is directed at, and what the specific outcome is. For example, Gregersen, Andersen, et al. (2023) illustrate that climate anger can stem from various sources - not all equally motivating for climate action. While some are angry about public apathy or lack of climate action, others explain their anger with skepticism towards the threat of climate change or dissatisfaction with mitigation measures. Consequently, feelings of anger might fuel activism aiming to pressure governments to act to address climate change, but also activism opposing climate solutions and policies, such as wind power development and carbon taxes. While the study supports that climate anger is related to climate change engagement, its effect depends on the specific outcome. While anger is particularly relevant for activism, other emotions, such as sadness or fear, are more relevant for policy support and individual mitigation behaviors. 
   
• Because emotional reactions indicate that a situation or phenomenon touches upon something people see as relevant, that they value and care about, both the presence and absence of emotion can be interesting. Gregersen, Doran, et al. (2023) investigated people’s explanations for (not) worrying about climate change. The most common reason for being at least somewhat worried was concern about the consequences of climate change, and those reporting high levels of worry were in particular more likely to bring up consequences for humans than those reporting medium worry. Respondent reporting low levels of worry referred to a broader range of reasons in their answers, such as believing in natural rather than human causes of climate change, expressing a sense of optimism towards potential solutions, or being discontent with political measures or public discourse on climate change. 

Recommendations for policy makers 

• Emotions related to climate change (sometimes referred to as ‘climate emotions’) are highly relevant for whether and how we succeed in mitigating climate change. Previous research has established important relationships between climate emotions on the one hand, and important outcomes such as risk perception, technology acceptance, policy support and mitigation- and adaptation behaviors on the other. Policymakers should consider the potentially motivating and (de)polarizing effect of climate emotions when developing communication strategies related to climate change mitigation.  
   
• Up until now, much of the research on climate emotions has been based on correlational data. This research has mainly focused on certain (individual-level) emotions such as worry and hope and has first and foremost investigated how these emotions relate to different forms of climate change engagement. Going forward, there is a clear need for expanding the focus to a broader range of both individual and collective emotions, further investigating social effects, and looking for causal relationships. 
   

References/Sources 

Gregersen, T., Andersen, G., & Tvinnereim, E. (2023). The strength and content of climate anger. Global environmental change, 82, 102738. https://doi.org/10.1016/j.gloenvcha.2023.102738 
 
Gregersen, T., Doran, R., & Storelv, S. (2023). Self-reported reasons for (not) being worried about climate change. Current Research in Ecological and Social Psychology, 5, 100154. https://doi.org/10.1016/j.cresp.2023.100154 

Author

Thea Gregersen, Senior Researcher, Health & Society, NORCE Norwegian Research Centre

No vote, but at least a voice? Opportunities and Barriers for young people's civic participation 

By Ida Vikøren Andersen, NORCE

Download policy brief in pdf

Summary 

This policy brief explores the opportunities and obstacles related to young people's civic participation. Drawing from research on young people's involvement in public debates about climate change, the brief highlights how children and youth can either wield or be denied rhetorical and political power in matters that significantly impact them.  
 
The policy brief recommends addressing the prevailing and enduring discourses that marginalise young voices in democratic discussions to foster a more inclusive and democratic society. This objective can be achieved through a concerted effort involving research, education, and critical reflection among key stakeholders, including educators, researchers, policymakers, and the media. 
 

Young climate protester, London 2019, Photo: Matt Harrop, licenced for reuse under cc-by-sa/2.0

Photo:

Matt Harrop

The issue

Climate change, with its profound and long-lasting impacts, particularly affects the young. But, due to their lack of voting rights, children and youth are excluded from influencing this issue through the ballot. 
 
Despite the high visibility of young climate activists like Greta Thunberg and the Fridays for Future movement, their voices often go unheard in the climate debate. Authorities, educators, and the media increasingly emphasise the importance of youth civic participation. Yet, the response to youth climate activism often undermines their democratic involvement.
 
In political debates, the young are frequently dismissed as inexperienced, irresponsible, and ill-informed instead of being recognised  as competent speakers with valid and valuable views. Increasingly, young climate activists are also demonised and portrayed as a threat to democracy in public discourse in many European countries.

Their civic participation does not threaten democracy, but their exclusion from democratic debates might. It might erode their trust in the potential for meaningful political discussions on pressing matters and diminish their faith in policymakers' willingness to prioritise their lives and the planet they will inherit, a sentiment echoed in the growing dissatisfaction among young people with how political leaders address climate-related issues.  

Key findings 

• Andersen (2023a) shows that while it has long been widely accepted that children and youth should have limited political influence, there is a growing movement advocating for increased political empowerment of young people. In several countries, debates are underway about lowering the voting age, and some European nations already grant 16-year-olds the right to vote. However, simply extending voting rights does not guarantee enhanced political representation and power for youth. Practical barriers, including the underrepresentation of young people among elected officials, can hinder them from having a voice in the political system, even if they have the right to vote. 
   
• Andersen (2023b) argues that acquiring political rights such as freedom of speech and the right to vote holds little value if not followed by the development of rhetorical citizenship. Rhetorical citizenship involves the ability to effectively articulate one’s opinions and be recognised  as a speaker with valid and valuable views. 
   
• Andersen (2023a) finds that a significant barrier to youth civic participation is the widespread discourses surrounding their roles in democracy. The view of children and youth as apolitical beings and immature citizens-in-the-making who are meant to play and learn rather than participate in the political sphere obstructs them from speaking and being heard. Their marginalisation is evident in how young people are far less frequently given a platform in the media, and when they do, they are often met with hateful comments or dismissed as immature, ill-informed, and irresponsible participants who are out of place in public debates. 
   
• Andersen (2023a; b) argues that efforts are needed in media and public discourse to recognise the valuable contributions of young voices and ensure that they are not dismissed or met with hateful rhetoric. This responsibility extends to researchers, policymakers, and media actors, prompting us to reflect on the assumptions about children and citizenship that underpin our actions. 

Recommendations for policy makers 

• Give 16-year-olds the right to vote and ensure representation of young people among elected officials.  
   
• Providing young people with training in rhetorical skills is crucial to empower them to assert their voices because, without the ability to express and argue for their views, their voices will remain unheard. 
   
• More research is needed to understand how young people participate in democracy and perceive their opportunities to do so. Existing research has largely focused on youth activists and those who manage to have their voices heard in the public sphere. However, it is equally important to investigate what prevents others from participating and making their voices heard. Such research should also address the prevailing discourses that marginalise young people in democratic debates. 

References 

Andersen, Ida V. (2023). Elever i streik – skoleskulkere eller systemkritiske medborgere? [Students on strike – shirkers or protesters?] Rhetorica Scandinavica, 27(86), 141-159. http://rhs.retorikforlaget.se/index.php/rhs/article/view/305  

Andersen, Ida V. (2023). Rhetorical citizenship and the environment. Climate Resilience and Sustainability, 2(2), e249. https://doi.org/10.1002/cli2.49 

Author

Ida Vikøren Andersen, NORCE Norwegian Research Centre  

Dealing with climate change in representative democracies

By Mari S. Helliesen, NORCE

Download policy brief in pdf

Summary

As one of the greatest global challenges of our time, climate change is also one of the greatest challenges facing democracies today. There are tensions in the relationship between representative democracy and climate change, including a lack of public support and government action, and the short-term horizon of electoral politics when facing the long-term climate crisis. There is a pressing need for climate action, while at the same time maintaining democratic ideals and functions, such as equal representation and responsiveness. Research on and discussions about how well-equipped representative democracy is to tackle the climate crisis are necessary going forward. 

The Issue

There is a tension between democracy and climate change in the sense that something needs to be done to tackle this global challenge, but there is a lack of public support and government action. Some argue that policy makers are reluctant to implement climate policies if they expect public opposition. Policies are implemented by governments who need the trust of their citizens. In addition, it has become increasingly difficult for governments to balance international commitments and promises to the electorate. In the context of climate change this is apparent in terms of nation states’ commitments through the United Nations Framework Convention on Climate Change (UNFCCC) and the current Paris Agreement, as well as EU climate policies and agreements, such as the EU Emissions Trading System (EU ETS). 

Another major challenge to necessary climate policy implementation is the short horizon of elected representatives and parties, operating in short electoral cycles. In limiting climate change and reducing emissions, democracy can by nature be moving too slowly and not be sufficiently efficient. Major effective policies are costly in the short-term, but its benefits will mostly be proven in a long-term. Furthermore, climate change is inherently long termed. Mitigation measures that will have the most effect will likely be costly. These more visible policies are difficult to get both citizens and policy makers behind. Still, research has shown that democracies do perform better than other regimes in terms of climate mitigation through cooperation in international environmental treaties, adopting stricter environmental policies, and curbing their CO2 emissions. 

When it comes to climate change, nation states alone will fall short – because the climate has no national borders. But democratic representation takes place within these national borders. This makes the relationship between states, on the one hand, and transnational and supranational organizations, on the other hand, in climate policy more complicated. At the same time, impatience from the general public and activist, communicated through protests, campaigns, and strikes, is directed both at national governments and the global community, for example the IPCC and yearly negotiation meetings between states (COP).  

Civil society is a vibrant and important part of democracy, and the relationship between organisations and protest activities and those in power can be beneficial. This way, people can be heard in different ways, and more often than just during elections. It is possible for citizens to put pressure on governments to act. We have witnessed multiple examples of this in recent years related to climate change, including Greta Thunberg and the Fridays for Future movement. There are also groups that take things even further with civil disobedience, such as Extinction Rebellion. 

My research

In my research on climate change perceptions and policy preferences, I utilize survey data from the Norwegian Citizen Panel and the Panel of Elected Representatives, with representatives at the local, regional, and national levels in Norway. This allows me both to compare responses from citizens and representatives on the same climate policy issues as well as to ask representatives directly about their views on dealing with climate change in a representative democracy. 

Key findings

• In a recent published article, I examined whether descriptively underrepresented groups also are substantively underrepresented on climate issues in Norway (Helliesen, 2023a). I found that the climate policy preferences of women and, especially, young are underrepresented by elected representatives This is the case even in the Norwegian context, with comparatively high levels of equality. Simultaneously, these two groups consistently support climate policies and call for climate action to a larger extent than their male and older counterparts. 
• In an ongoing study (Helliesen, 2023b), I ask elected representatives about their perceptions of dealing with the climate crisis within democracies. Preliminary results show that a majority of the Norwegian representatives in the Panel of Elected Representatives believe that the climate crisis can be solved through regular decision making. This belief is stronger for men than women, and less strong for politicians in the radical right Progress Party (FrP) than other parties. The belief also increases when politicians worry more about climate change and place themselves more to the left on the political left-right scale. 
• A majority of the surveyed politicians oppose people engaging in civil disobedience to pressure governments for climate action. Support for civil disobedience is higher for women than men, and for representatives in the Green (MDG), Red (Rødt), and Socialist Left (SV) parties. Support for civil disobedience among elected representatives also increases with higher education, climate worry, and placement to the left on the political spectrum (Helliesen, 2023b). 

Recommendations

• Ensure that women and young people are better represented in formal politics. Especially youth, who participate less in formal political channels such as voting and running for office, but more so in unconventional forms such as protests and civil disobedience
• There is need for more research and discussions on how well-equipped representative democracy is to tackle the climate crisis.

References

Helliesen, M. S. (2023a). Unequal Representation of Women and Youth on Climate Policy Issues. Representation, 59(4), 615-632. https://doi.org/10.1080/00344893.2022.2133001

Helliesen, M. S. (2023b). Representative democracy and climate change: perceptions and preferences of elected representatives. [Unpublished paper].

Author

Mari S. Helliesen, Senior Researcher, NORCE

Enhancing public sector AI: Empowering citizens through informed deliberation

By Sveinung Arnesen, NORCE

Download policy brief in pdf

Summary

The rising use of artificial intelligence (AI) and machine learning in the public sector, presents unique challenges in terms of discrimination, fairness, and transparency in AI decision-making. Addressing these challenges requires not only a deep understanding of AI but also an informed and engaged citizenry, especially in democratic societies where public sector AI must adhere to democratic principles.

Recognizing the critical role of public opinion in legitimizing AI systems, our approach involves a Deliberative Poll, engaging citizens directly in the conversation about AI in the public sector. 
Participants who engaged in the Deliberative Poll reported a significant rise in their understanding of artificial intelligence and its specific applications. This heightened awareness correlated with a more favorable attitude towards AI’s role in crucial public sector decisions.

This research underscores the importance of informed citizen participation in shaping AI policies and the potential of deliberative events as a tool for democratically integrating AI into public administration.

Background and challenges

The use of artificial intelligence is on the rise in the public sector. A report published by the European Commission in 2022 identified 686 public sector AI use cases in its member states plus some other European countries, most of which were based on machine learning (Noordt et al. 2022). The use cases are growing rapidly. 

These computational advances combined with the growing availability of data raise novel governance challenges with respect to discrimination, fairness, and transparency in AI decision-making.

In democracies, input from citizens is critical to developing legitimate AI systems, not least in the public sector which has obligations to make sure its use of AI adheres to democratic principles. Yet, in these early stages of AI implementation, citizens have little experience and knowledge about AI. A pressing concerns is thus how to put society in-the-loop when developing and implementing AI tools (Rawhan 2018).

AI in the Norwegian Public Sector

While AI is not yet operational in decisions directly impacting individuals, its potential use is being explored. One such area is within the Norwegian Labour and Welfare Administration (NAV), aiming to predict sick leave durations to facilitate targeted support for returning to work. Currently, NAV manually selects individuals for dialogue meetings about sick leave, but AI models could refine this process by predicting sick leave lengths, aiding in decision-making.

Similarly, the allocation of refugees to municipalities, currently a manual process by The Directorate of Integration and Diversity (IMDi), could be enhanced by AI. Case managers use various refugee data to determine suitable settlements, a process that might be streamlined and improved with AI.

Parole decisions, another critical area, currently rely on case managers’ assessments. While Norway has not adopted AI for this, other countries use recidivism prediction models, though they are not without controversy.

My research

Our research aimed to enhance public understanding of AI through experimental education and then reassess their attitudes (Arnesen et al. 2023).

We recruited 207 residents of Norway for a Deliberative Poll, utilizing the national population registry for random selection. This method ensured a representative and high-quality sample. Participants were divided into two groups, discussing either CO2 capture and storage or AI in the public sector. 

Prior to the event, participants received comprehensive briefing materials about their respective topics. On the deliberation day, they used an online deliberation platform developed by our partners at Stanford University for structured discussions. Guided by an automated moderator, they discussed policy proposals, formulated questions for experts, and participated in plenary sessions with expert responses. The event, lasting five hours, concluded with the same survey they completed earlier to assess changes in opinions and understanding.

Key findings

• Participants in the treatment group reported a significant increase in knowledge about AI after the deliberation event, compared with the control group. Their knowledge about the specific tasks where AI potentially can be used also increased. 
   
• With the increase in knowledge, the participants in the treatment group also became more positive towards the use of AI when making decisions in these areas.

Photo:

NORCE

Policy recommendations

• To maintain legitimacy and trust among citizens, authorities need to involve citizens in deliberations about how and where AI should be used. 
   
• We thus advocate for the integration of deliberative events in AI policy development at national as well as the EU level. Participants should be representative of the citizenry and be given the resources needed to make an informed opinion on the questions at hand. Organizers should strive to present balanced information material and ensure diverse expert involvement.

References/Sources

Arnesen, Sveinung, T.S. Broderstad, J.F. Fishkin, M.P. Johannesson, A. Siu. 2023. Knowledge and Support for AI in the Public Sector: A Deliberative Poll Experiment. Conference paper presented at the Annual Conference of the Southern Political Science Association, St. Pete’s Beach, USA.

Noordt, Tangi L. van, M. Combetto, D. Gattwinkel, F. Pignatelli, and AI Watch. 2022. Euro-
pean Landscape on the Use of Artificial Intelligence by the Public Sector. Technical report.
Publications Office of the European Union. https://doi.org/10:2760/39336.

Rahwan, Iyad. 2018. “Society-in-the-loop: programming the algorithmic social contract.” Ethics and information technology 20 (1): 5–14.

Author

Sveinung Arnesen, NORCE

AI in educational settings and data protection concerns

By Malgorzata Cyndecka, University of Bergen

Download policy brief in pdf

Summary

While AI-based tools and applications are being promoted as a game changer for education, there is not sufficient awareness of their implications to children’s right to data protection. With the General Data Protection Regulation being rather disappointing as regards protecting minors, and the proposal for AI Act that (at least currently) merely qualifying AI in education as high-risk AI systems, it is time to take children’s right to data protection more seriously.

Photo:

Generated with Dall-E 3

The issue

While there is no consensus amongst the experts as regards the definition of AI, there is little doubt that Al-based tools and applications are increasingly having an impact on education. AI brings opportunities as well as numerous challenges and threats to pupils’ fundamental rights and freedoms, including their right to data protection.

Although “children merit specific protection with regard to their personal data” (GDPR recital 38), those data are far too often uncritically shared with or sold to commercial providers of AI-based or even merely apparently AI-based applications promising prodigious results. Those results, however, have been rarely demonstrated while one risks a redistribution of authority. It will no longer be schools or public authorities that decide on curriculum, but providers of AI-applications; it will no longer be a teacher who decides on learning activities in a classroom, but an algorithm.

Moreover, lack of transparency, bias, discrimination, and exclusion not only threaten children’s data protection, but also undermine the foundations of democracy such as openness and the possibility of making informed choices. The research on AI in education with regard to such issues as transparency and explainability is scarce.

Finally, as shown by amongst others the results of the EdTech “AI Sandbox” of the French Data Protection Authority, CNIL, there is in general little knowledge on how to comply with data protection regulations when implementing EdTech in schools.

My research

In my research on privacy and data protection, I focus on researching the consequences of deploying new technologies, such as AI, as well as on regulating AI, in particular in the sector of education. As a member of the Council of Europe’s Expert Group “Artificial Intelligence in Education”, I am involved in working on a proposal for a legal instrument that will regulate using AI in educational settings.

In 2021/2022 I was involved in a project “Activity data for assessment and adaptation” (AVT) where pupils’ personal data were used for the purposes of learning analytics. The project was led by the Norwegian Association of Local and Regional Authorities (KS), the Centre for the Science of Learning & Technology (SLATE) at the University of Bergen (UiB) and the City of Oslo's Education Agency. As data protection expert, I have analysed amongst others the legal basis for processing of pupils’ personal data with the view of identifying both the legal basis in the GDPR and the required supplementary legal basis in the national legislation (opplæringslova). The AVT project was selected by the Norwegian data protection authority, Datatilsynet, to participate in its first edition of “Sandbox for responsible AI”.

Currently, I am involved in “EduTrust AI”, a research project led by SLATE and funded by The Trond Mohn Foundation. The main goals of the project are: a) to identify layers of trust associated with the use of AI in the educational sector that considers the complex accountability relationships, b) to develop new knowledge, methods, guidelines, and tools for more reliable AI systems, c) to translate insights about legal, psychological and sociocultural determinants of trust into legal requirements, and d) to provide input for practicable frameworks related to the challenging questions surrounding the use of student data and AI systems in education.

Results and recommendations

EU policy makers must consider regulating the deployment of AI in educational settings. While the GDPR is technology neutral and does not specifically address AI, the current proposal of AI Act does not specifically regulate AI in education. Also, a general statement on that the AI Act does not prejudice the GDPR does not provide any clarifications on their complex and critical relation and thus leaves scope for gaps and legal uncertainty.

As the GDPR disappoints in terms of providing an appropriate protection of children’s right to data protection, EU policy makers must consider how this issue may be addressed at the EU level. Although a revision of the GDPR would be an immense task, one cannot ignore the challenges and threats raised by AI as they may profoundly affect Europe’s next generations and its future.

References/Sources

The AVT project – exit report from the “Sandbox for responsible AI”, https://www.datatilsynet.no/en/regulations-and-tools/sandbox-for-artificial-intelligence/reports/avt--exit-report/ 

Malgorzata Cyndecka, «Bruk og misbruk av personopplysninger om egne barn i sosiale medier i lys av barns rett til personvern i barnekonvensjonen artikkel 16» («Use and misue of one’s own children’s personal data in social media in light of children’s right to data protection and Article 16 of the Convention of the Rights of a Child») in:  Fornes, I., Nylund A. Sperr A.K, “Barnekonvensjonen i norsk rett. Prinsipper og praksis” (“The Convention of the Rights of a Child in the Norwegian law. Principles and practice”), Gyldendal 2023, pp. 113-137.

Author

Associate professor Malgorzata Cyndecka, PhD, Faculty of Law, University of Bergen
Affiliated with the Center for the Science of Learning and Technology (SLATE), University of Bergen

Handshaking controversies 

By Elisabeth Ivarsflaten, University of Bergen

Download policy brief in pdf

Background

In the last decades we have seen a number of controversies over handshaking in European democracies that involve Muslims. Many Muslims are happy to shake hands, but some observant Muslims believe it is wrong to touch members of the opposite sex who are not relatives. The question is how handshake refusals should be handled in liberal democracies.

Abstractly formulated, the logic of conflict over gender and handshaking is strict. The traditionally observant Muslim is asked to do what their convictions condemn as wrong, shaking the hands of a member of the opposite sex. The non-Muslim majority is asked to accept what many of them condemn as wrong, treating women differently and worse. 

When confronted with this conflict, public authorities in Europe have often insisted on conformity with the handshaking practice. The president of France refused citizenship to a Muslim woman who would not shake hands with a male official at a citizenship ceremony.  In December 2018, Denmark passed a law requiring new citizens to shake hands at their naturalization ceremony.  The Administrative Court of Baden-Württemberg rejected a citizenship application from a male Muslim, because he refused to shake hands with a female immigration official.  In Switzerland, Muslim students were forced to shake hands with teachers of the opposite sex.  In Norway, there was recently a similar controversy over handshake refusals in a school setting. Muslim job applicants in both the Netherlands and Sweden were turned down after handshake refusals.

The study

In a study funded by the European Research Council (ERC) and forthcoming in the British Journal of Political Science, a group of European and American researchers led from the University of Bergen, argues that introducing a substitute gesture of respect—putting the hand on the heart—is a viable alternative solution. 

In the study, the team experimentally tests responses in a representative sample of German citizens in three common handshake-refusal situations. Half the time, the substitute gesture is introduced and half the time, it is not.
The study finds a remarkably strong and consistent effect of introducing the alternative gesture of respect: Insistence on conformity with the handshaking practice drops notably and consistently.  Three quarters of respondents think Muslims should be asked to conform to the handshaking custom if they refuse to shake hands with female politicians (76 per cent), female teachers (70 per cent), and female HR managers (74 per cent).  When the substitute gesture is introduced, insistence on conformity drops by 20 percentage points or more.  If the gender of the politicians, teachers, and HR managers is left unspecified, only around 40 per cent of the non-Muslim public insists on conformity to the practice of handshaking after the alternative gesture of respect has been offered.

More in-depth analysis of the results shows an important limitation: The alternative gesture of respect reduces insistence on conformity in most segments of the population, but not among the least tolerant citizens and those who place themselves on the far-right politically. 

Specific policy recommendation

Public authorities in Europe should not insist on handshaking. They should instead insist on gestures to signal respect. Alternative gestures of respect should be allowed.

General policy recommendation 

In the foreseeable future, European societies that honor pluralism will have to live amicably with religiously grounded differences. Thanks to the ingenuity and versatility of cultural customs to signal respect, value conflicts can be open to resolution in everyday encounters without minorities or majorities having to forsake their convictions.  Requiring that minorities, Muslims or others, conform should be avoided if alternative solutions acceptable to both exist or can be developed.

The study

Ivarsflaten, Elisabeth, Marc Helbling, Paul M. Sniderman, and Richard Traunmüller (forthcoming), “Value Conflicts Revisited: Muslims, Gender Equality, and Gestures of Respect.” British Journal of Political Science.

Authors

Elisabeth Ivarsflaten is Professor at the Department of Government, the University of Bergen.
Marc Helbling is Professor at the Faculty of the Social Sciences, the University of Mannheim.
Paul M. Sniderman is the Fairleigh S. Dickinson Jr. Professor in Public Policy at the Department of Political Science, Stanford University.
Richard Traunmüller is Professor at the Faculty of the Social Sciences, the University of Mannheim.

The research was funded by the ERC Consolidator Grant #101001133 INCLUDE.

From clicks to civic engagement

By Andreas Nordang Uhre, NORCE

Download policy brief in pdf

Summary

The digitalisation of political discourse – commonly exemplified by the growth of social media platforms – is frequently depicted as a significant threat to democracy. Yet, it also carries the potential to engage traditionally marginalised groups and individuals, integrating them into the democratic process. My research suggests that the digitalisation of political discourse can help lower the threshold for political engagement, by mobilising groups or individual citizens who might otherwise have found the bar for joining formal politics too high. To harness this potential, decision-makers should facilitate citizens’ entry into formal political arenas. 

The Issue

Is democracy in crisis? Participation in political activities is dropping, especially among young people. Trust in political institutions is fading, and growing inequality feeds distrust and cynicism. The brunt of the blame for these developments is laid at the door of digitalization of political discourse, in particular the rise of social media platforms. They are accused of contributing to the deterioration of the democratic landscape through the dissemination of fake news, questionable science, computational propaganda, aggressive microtargeting, and political advertising. This, in turn, fosters increasingly confined ‘echo chambers’ of personalised news and connections with like-minded individuals. This restricted exposure reinforces similar ideological viewpoints, creating a feedback loop of opinions.

However, the emergence of social media platforms has undeniably also created spaces that enable new and diverse groups of citizens, including those traditionally marginalised, to participate in political discourse (Margetts 2018). Two decades ago, engaging in politics typically required joining a political party, participating in organised interest groups, attending meetings, or canvassing door-to-door. For many individuals, the associated costs in terms of time, effort, and resources were often deemed prohibitive. Consequently, politics became predominantly the realm of an activist elite. In contrast, contemporary dynamics have witnessed the emergence of small, manageable actions that are attracting a new demographic into politics. Notably, this trend is particularly evident among young people, whose longstanding absence from political participation has been lamented by commentators for years. 

Lance Bennet and Alexandra Segerberg (2012) introduced the concept connective action. They posited that societal and cultural shifts associated with globalisation and individualisation have led to disruptions in group affiliations and institutional allegiances. Additionally, these changes have fostered individualised perspectives, prompting political engagement as a manifestation of personal aspirations, lifestyles, and grievances. Bennet and Segerberg assert that as communication increasingly takes place within digitised decentralised networks, opportunities for decentralised and self-organised collective action emerge. The defining feature of these networks is their ability to operate through social media channels without requiring strong organisational control or the establishment of symbolic communities, which distinguishes them from traditional forms of collective action that rely on forma organisational structures. This gives these networks a more fluid, short-lived, and ad-hoc character than traditional forms of collective action. 

Bernard Eljoras and Ivar Eimhjellen (2019) acknowledge the widespread adoption of digital communication technology by both individuals and organisations. Yet, they contend that while the impact of these digital networks on democratisation and mobilisation is positive, they are limited to the initial phase of the political engagement cycle. Moreover, they argue that for digitally organised networks to endure in the long term, there is a need for formalisation and organisational structure. 

My research

During the summer of 2023, I conducted interviews with the founders of Motvind, a Norwegian anti-wind power interest group founded in 2019, and the anti-road tolls political party FNB, founded in 2014.    The objective was to understand the motivations that led these citizens to venture into the realm of political entrepreneurship. 

Key findings

• One common thread between the two organisations was the founders’ initial political apathy. Their paths into activism shared a common narrative: upon learning about plans to install wind turbines or new road tolls in localities where they lived, hiked, or grew up, they turned to Facebook as a tool for finding and sharing information. 
   
• Facebook also served as a means to connect with like-minded citizens and, eventually, organise. Motvind has achieved significant success by placing their criticisms of wind power on the political agenda, persuading elected officials to impede local developments. Similarly, the FNB caused a minor political earthquake in the 2019 Norwegian local elections, securing substantial representation in various local assemblies. Particularly noteworthy was their impressive 16.8 % of the vote in Bergen, Norway’s second-largest city.
   
• One of the informants recounted her journey from complete political apathy, via active involvement with Motvind, to ultimately running for mayor in 2023 as a “total personal transformation”. This involved a comprehensive shift in her media consumption, a much-broadened personal network, and a newfound interest in civic issues. The informants from both organisations emphasized that without the presence of Facebook or similar platforms, their groups would never have come into existence. The initial connective action played a vital role in their political mobilisation. 
   
• However, to attain political influence, they recognised the need to formalise, organise, and transition into real-world activities, essentially shifting from connective to collective action. These transitions were significantly facilitated by Norway’s openness to interest groups, its multiparty electoral system, and the absence of electoral thresholds at the local level of governance. 

Recommendations for European policy makers

•    Improve framework conditions for NGOs.
•    Advocate for multiparty electoral systems throughout the European Union.
•    Eliminate electoral thresholds in local politics. 

References

Bennett, W. Lance and Alexandra Segerberg. 2012. The Logic of Connective Action. Information, Communication & Society 15(5): 739-768. https://www.tandfonline.com/doi/full/10.1080/1369118X.2012.670661

Enjolras, Bernard and Ivar Eimhjellen (eds.). 2019. Fra kollektiv til konnektiv handling. Nye former for samfunnsengasjement og kollektiv handling. Oslo: Cappelen Damm Akademisk. https://press.nordicopenaccess.no/index.php/noasp/catalog/book/45

Margetts, Helen. 2018. Rethinking Democracy With Social Media. The Political Quarterly 90: 107-123. https://onlinelibrary.wiley.com/doi/10.1111/1467-923X.12574

Author

Andreas Nordang Uhre, Senior Researcher, NORCE

Generative AI poses a risk to European culture 

By Jill Walker Rettberg, Center for Digital Narrative, University of Bergen 

Download policy brief in pdf

The issue

We risk losing intangible European cultural heritage if the generative AI of the future is not trained on European data.  

Generative AI produces texts and images that are statistically probable based on the data the model was trained on. This means that minority viewpoints are less likely to be reflected in AI-generated content, which could lead to a tyranny of the majority. This is a problem for democracy in general, because of the potential loss of diverse voices. It is a problem for Europe in particular because European content is underrepresented in popular generative AI models like ChatGPT. 

Policy recommendations for Europe

• Support the development of national and European AI models.  
• Work to increase European content in international and commercial AI models while supporting creators’ rights to their content: 
  • Develop legal frameworks for collective rights so copyrighted material can be included in training data while creators are compensated and their intellectual property rights respected. 
  • Make European data that is not copyrighted available as open access datasets that can be used to train AI models 
  • Prioritise making data from small and minority languages accessible in collaboration with language communities. 
• Increase the general public’s understanding of how generative AI works so we are better equipped to use it productively without succumbing to its biases and flaws. 

Background

Generative AI has gone mainstream since the introduction of ChatGPT in November 2022. In sectors and industries from education and media production to sales and marketing people are using AI to create, edit or inspire new texts and images.  

European policy-makers need to be aware that most generative AI models are trained on English language content, predominantly from the United States. Although models like OpenAI’s ChatGPT, Google’s Bard and Microsoft’s Bing Copilot can produce texts in many languages, their base language is English. This means that Anglo-American genres, styles and ideas are the basis for the texts and images produced by AI models like ChatGPT. 

This leads to new types of AI bias that threaten European culture. Previous discussions of AI bias often point to bias in the training data. For example, a facial recognition system trained mostly on images of white men’s’ faces will recognize white men better than it recognizes black women. Generative AI is known to have biases as well. For example, an AI model given the prompt “terrorist” will tend to produce pictures of people who look Arabic, while the prompt “nurse” produces pictures of women. 

The biases that may lead to the loss of European cultural heritage are more subtle. Generative AI is trained on stories, and stories are culturally specific. For example, in Norway Thorbjørn Egner’s Folk og røvere i Kardemomme By (Folk and Robbers in Cardamom Town) is a well-known children’s book and musical that features three comical robbers who steal food because they are hungry and don’t understand that work is necessary. After being caught stealing sausages and chocolate they are rehabilitated by the kind police officer and townsfolk and end up saving the town from a fire. This story is not just a shared cultural reference, it is a cultural support for the Norwegian criminal justice system’s focus on rehabilitation above punishment. Hollywood stories about robbers or criminals are very different: Bank heist and gangster movies glorify criminals, while Disney movies have unambiguous villains who die at the end of the movie.  

An AI model trained on stories about bank heists or villains will produce stories that are similar to what it was trained on. What do we miss if generative AI never produces stories about criminals with rehabilitation as a solution? 

Europe is culturally and linguistically diverse. Different countries, regions and communities have stories that are important to their identity.  

If we fail to develop generative AI that supports European culture, we risk losing our cultural heritage. We risk losing our stories. 

Hindrances to generative AI that supports European culture are: 

• Commercial – the dominant companies (OpenAI, Google, Microsoft, Baidu) are based in the USA or China and do not have incentives to use European training data.  
   
• Legal – high quality training data is usually copyrighted. Data scraped from the internet may contain personal information (e.g. photos of individuals, social media posts). 
   
• Infrastructural – we need to increase capacity for research and development in Europe 
   
• Cultural – human-centred generative AI requires collaboration between the humanities and computer science. Statistics and computer science were developed for numeric data. When the training data for and output from AI models are cultural (language, images and records of human behaviour) we need input from other fields. 

Research 

I have researched AI bias in the ERC funded project Machine Vision in Everyday Life: Playful Interactions with Visual Technologies in Digital Art, Games, Narratives and Social Media (2018-2024), finding that although biases can be encoded in technology, the cultural contexts in which technologies are introduced are just as significant. Generative AI complicates this because the training data that shapes the AI model is from a specific cultural context. 

In July 2023 we launched the Center for Digital Narrative (CDN), a Norwegian Centre for Research Excellence at the University of Bergen that is funded by the Norwegian Research Council for ten years. Scott Rettberg and I are co-directors of the CDN, with Scott Rettberg leading it for the first five years.  

The CDN focuses on digital narratives ranging from video games, electronic literature, and social media narratives to AI-generated narratives and other emerging genres. Generative AI a central concern for the CDN. 

References 

Rettberg, Jill Walker. Machine Vision: How Algorithms are Changing the Way We See the World. Cambridge: Polity Press, 2023. 

Rettberg, Scott, Talan Memmott, Jill Walker Rettberg, Jason Nelson, and Patrick Lichty. «AIwriting: Relations Between Image Generation and Digital Writing», 2023. https://doi.org/10.48550/ARXIV.2305.10834. 

Author 

Jill Walker Rettberg is Professor of Digital Culture and Co-Director of the Center for Digital Narrative at the University of Bergen. 

This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771800), and from the Research Council of Norway through its Centres of Excellence scheme, project number 332643.