Project summary

In 2025, 37.9% of people aged 16 to 74 in Spain reported using generative AI. At the same time, Spain is moving from talking about AI regulation to actually implementing it, including the EU AI Act alongside its national AI sandbox and other governance arrangements.

There is already plenty of polling asking whether people feel optimistic, worried or indifferent about AI. We want to look at something slightly different: what makes the use and governance of AI feel legitimate to different parts of the public?

More in Common España would run a national pilot with a representative survey of 2,000 adults and a smaller piece of qualitative research. We would analyse the findings through our existing Spanish population segmentation.

The research would look at how people understand AI now, including the difference between the tools they already use and more advanced systems. We also want to know which risks they actually take seriously, who they think should be responsible when things go wrong, and what kinds of safeguards seem reasonable rather than excessive or meaningless.

Part of the survey would test different ways of explaining AI risks, responsibility and safeguards. We are not just interested in finding the wording that produces the most support. We want to know what people understand, what they believe, what they are sceptical about and why.

We are particularly interested in whether attitudes to AI fall along the usual political or demographic lines. They may not. Our segmentation gives us another way of looking at this, based on people's underlying values and worldviews.

The findings would be public and shared with people working on AI governance in Spain and Europe. This is a pilot. If it produces something useful, the same approach could be tested in other More in Common countries later.

What are this project's goals? How will you achieve them?

The main goal is fairly simple: give people working on AI governance a better idea of what the public actually thinks about some of the questions they are working on.

That includes policymakers, regulators, companies, researchers and civil society.

We want to understand:

• how people in Spain currently use and understand AI;

• whether people distinguish between everyday AI tools and more advanced AI systems;

• which risks people find credible and which seem remote, confusing or exaggerated;

• how trust in institutions, personal agency and ideas about responsibility shape people's views;

• who people think should be responsible for overseeing AI;

• how people react to specific safeguards, for example independent testing, transparency requirements, human oversight or limits on certain uses;

• where views differ between More in Common's Spanish population segments;

• and, just as importantly, where people with very different worldviews agree.

Before finalising the research questions, we would speak to a small group of people working on AI governance and policy. We do not want to produce another general survey on AI if the people working in this area already have the information they need. We would ask them where the gaps are and what public opinion evidence would actually be useful.

The research itself would have three main parts.

A nationally representative survey of 2,000 adults in Spain

The survey would include More in Common España's population segmentation. This means we can look at people's responses through their values and worldviews, alongside the usual demographic and political categories.

Experimental testing

Some respondents would see different explanations of AI risks, responsibilities and possible safeguards. We would compare how people understand them, whether they find them credible, whether they seem proportionate, and how they affect support.

Qualitative research

We would also run focus groups. These would help us understand why people are responding in certain ways and explore questions that are difficult to answer properly through polling alone. We would use our segmentation model when choosing participants.

The final output would be a short public report rather than a very long piece of research. We would pull out the findings that seem genuinely useful and explain where different parts of the public are coming from.

We would also offer briefings to organisations, researchers and other people working on AI governance in Spain and at EU level.

How will this funding be used?

We are seeking around 65,000 USD for the full pilot.

The project can be scaled depending on the amount raised.

Most of the funding would go towards survey fieldwork, focus groups, research design, analysis, staff time and project management. There would also be some costs connected to publishing the findings and sharing them with relevant organisations.

We think the minimum useful version of the project would cost around 45,000-50,000 USD.

At that level, we could still run the representative survey, analyse it using our segmentation and publish the findings. We would probably have to reduce the qualitative research and some of the follow-up engagement.

At 65,000 USD, we could include the focus groups, a stronger experimental testing component and more direct work with people and organisations working on AI governance.

Who is on your team? What's your track record on similar projects?

More in Common España has been conducting public opinion research in Spain since 2021, with a permanent Spanish team in place since 2024. Our work combines nationally representative polling, qualitative research and population segmentation, with a particular focus on understanding why different groups think the way they do and where apparent political divisions do or do not reflect wider public opinion.

One of our main pieces of work is the Atlas de la Polarización, published in 2025. The Atlas maps the main areas of political and social division in Spain, what appears to be driving them, and areas where there is considerably more agreement than the public debate might suggest. This is particularly relevant to the proposed AI project because we would use a similar approach: rather than treating “the Spanish public” as one group, we want to understand which attitudes towards AI are genuinely polarised, which divide people in less obvious ways, and where there may be broad agreement across groups with very different values or politics.

More in Common España has also applied this kind of public opinion research to climate and energy, migration, housing, media use, international affairs and other contested issues. Recent work includes La sociedad española y la transición ecológica, which included survey experiments alongside broader research into public attitudes towards climate and energy, and Después del apagón, a rapid-response survey carried out immediately after Spain's April 2025 blackout. We have also run research on immigration, including Spain's first More in Common study specifically focused on public attitudes towards migration.

The Spanish team has also developed its own electoral polling model, Pulso electoral, with its methodology and survey microdata made publicly available. More broadly, More in Common España makes questionnaires and microdata from its studies available on request and has published more than ten studies covering a wide range of public issues.

Our previous Spanish studies can be found here:
https://moreincommon.es/estudios.html

The core team for this project would include:

• Luis Aguado Alba, Director of More in Common España and project lead

• Tarek Jaziri-Arjona, Researcher, leading research and survey design

• a second Research Associate supporting qualitative research and segmentation

• relevant colleagues from More in Common's international research team where useful

We have not applied this approach to AI governance in Spain before. That is why we are proposing it as a pilot rather than a larger programme.

Our expertise is public opinion rather than technical AI research. We would involve people working directly on AI governance when developing the survey and ask external experts to review the technical descriptions and scenarios used in the research.

What are the most likely causes and outcomes if this project fails?

One possibility is that people simply do not have very developed views about advanced AI yet.

If that is the case, asking them very detailed questions could produce answers that look more meaningful than they really are. We would try to avoid that by using short, neutral explanations and testing the language before fieldwork. The focus groups would also help us see when people have a real view and when they are reacting to an unfamiliar idea for the first time.

Another possibility is that we produce interesting findings that are not especially useful to people working on AI governance.

That is why we want to speak to some of those people before designing the research, rather than only showing them the results afterwards.

It is also possible that our segmentation does not tell us much more about attitudes to AI than standard demographic or political categories do.

We do not know yet.

If that is what we find, there would probably be less reason to repeat this project across several countries. But that would still be useful to know from a relatively small pilot before putting substantially more money into a multi-country version.

Even if the pilot does not lead to further work, the survey data and main findings would still be made public.

How much money have you raised in the last 12 months, and from where?

We have just begun seeking funding for this project and are yet to raise funding to this specific project. Our usual sources of income are philanthropic foundations and grants supporting public opinion research towards social cohesion, we are yet to find a funder interested enough to support the project but hope to continue building the appeal across Spain, Europe and beyond.