FOIs about AI Existential and Catastrophic Risks

Brussels

Seeking Fundingactive6 people

Submitting FOI requests across EU Member States to reveal how governments understand and address advanced AI risks, creating evidence for accountability, advocacy and stronger policy.

Endorsed by

Ryan

Marcus+6

Submitting FOI requests across EU Member States to reveal how governments understand and address advanced AI risks, creating evidence for accountability, advocacy and stronger policy.

Endorsed by

Ryan

Marcus+6

Project Details

We have drafted an initial Freedom of Information (FOI) request concerning advanced AI risks, addressed to EU Member States. We are currently updating it following review by the Future of Life Institute team, and it will be finalised after a second review by a technology law professor. The draft is included in the private information section.
The FOI requests will be submitted in several EU Member States. We will report on the responses, analyse the findings, and provide comparative insights into how different governments understand and address catastrophic and existential AI risks.
Governments often avoid publicly answering questions about these risks. The requests will create a public record of governments’ awareness, internal thinking and actions taken.
Findings may support public accountability, inform future legal and policy work, and provide civil society with evidence for more informed and targeted advocacy.
Here's an introduction to our organisation.
Law for AI Safety Team:
- Raphaël Weuts: EU AI law and governance expert; author of On Accuracy of European AI Law (2026); advised the Belgian government and MEPs on the EU AI Act/GPAI Code; developed INASTI-RSVZ’s framework; led medical AI projects; former visiting professor in AI.
- Karolina Gruzel: Successif's AI Policy Strategy Fellow, successfully managed cyber attribution legal project; founded Moral Ambition for AI; op-eds (one, two, three), supported GAIGANow development and communications; Educational background in European and International Law and PPE.
- Cristian Teodorescu: Senior government/industry/ngo professional; awarded multiple times for humanitarian work; internal distinctions for inter-agency cooperation with the EU; granted off-cycle promotion for resilient and effective leadership on high-impact international affairs projects; given the highest-level national security clearance of an EU member state; Educational background in Human Rights Law.
- Raluca Spataru: Romanian lawyer with 15 years’ experience.
- Didier Coeurnelle: AI governance advocate, co-organised PauseCon Brussels, ordered a global ASI attitudes poll, speaker at WSIS Forum 2026 and LSAIR 2026, published author and active in civil society, including the Steering Committee of Global AI Governance Alliance; Belgian lawyer.
- Anonymous (see private information) - Senior Ops and Finance professional.
- Two volunteers, Julia (graphic designer) and Harry (website designer)

Theory of Impact

Activities

Finalise and submit FOI requests on catastrophic and existential AI risks in several EU Member States
Analyse and compare the responses
Publish findings and communicate them to the public, policymakers, and civil society.

Outputs

Government responses and non-responses collected
Comparative report on awareness, approaches, preparedness, and action
Public record of how governments are thinking about and addressing advanced AI risks

Interim outcomes

Better targeted national advocacy
Better governmental responses to x-risks questions due to visibility and public pressure
Greater transparency about government awareness and preparedness
Gaps, inconsistencies, and promising approaches identified
Increased public and political pressure on governments that are not adequately addressing the risks

Long-term outcomes

People

Karolina Gruzel

Team Member

Track Record

We have drafted an initial Freedom of Information (FOI) request related to advanced AI risks, aimed at Member States, and are currently finalising it following review by the Future of Life Institute team. Once we are established, the FOI requests will be submitted in a number of Member States, and we plan to report on the outcomes.
We have begun building our network. We have established a relationship with PauseAI and are working with them to update their onboarding process so that volunteers with legal backgrounds can be directed to us. Through attending conferences, we have connected with several experts in the field, including individuals involved with the EU AI Office, who have expressed a willingness to stay in touch and be informed about future opportunities for collaboration.
We have secured pro bono collaboration from a global law firm, which will help us register our NGO in Brussels and is open to supporting us further, for example by contributing some hours to our research or connecting us with lawyers interested in AI risks.
Our Board member, Didier Cournelle, introduced Law for AI Safety’s mission at the 2nd International Conference on Large-Scale AI Risks.
On 25 June, three members of our team — Raphaël Weuts, Karolina Gruzel, and Cristian Teodorescu — participated in the “AI Act in Romania: From Regulation to Implementation” conference at the European Parliament in Brussels. The event was hosted by Maria Grapini, Vice-Chair of the Committee on the Internal Market and Consumer Protection, and attended by, among others, Victor Negrescu, Vice-President of the European Parliament; Brando Benifei, former co-rapporteur on the EU AI Act; Martin Ulbrich, AI Policy Officer at the European Commission; and representatives of Romanian industry.

Grants Received

Add grant

grantmaking.ai Launch Round grant

from grantmaking.ai

$20,000

Discussion

Private comments in this thread are only shown to approved funders and grant reviewers. Email hi@grantmaking.ai or to request access.

Private UserPrivate1mo?

Genevieve SheaEndorsement1mo

Great idea and a very well written and organized proposal. FOI are a great resource for insight into how governments are actually thinking about catastrophic AI risks and where stronger engagement is needed. Comparative evidence across Member States could also become a useful resource for other projects fundraising on this site.

Karolina Gruzel1mo

Thank you, Genevieve!

I agree that the data could be useful for many projects, including ones focused on strategic advocacy.

Pim KangEndorsement1mo

I believe that we should have more advocacy in AI safety at this moment and this freedom of information route seems like one good way to start.

Karolina Gruzel1mo

Thanks, Pim! We also believe that the insights could support the advocacy efforts of other orgs

Olin Thakur1mo

I find this project very interesting as it has the potential to promote transparency and accountability alongside preparedness . Most discussions in AI Governance still overlook the ''accountability'' part, who should be held accountable when things go wrong.

Karolina Gruzel1mo

Thanks, Olin, for highlighting the importance of transparency and accountability. In the case of FOIs, accountability can take place for two reasons: insufficient or lacking responses, or insufficient actions.

AAbeer SharmaEndorsement1mo

This seems like a fairly straightforward project that’s easy and clear to implement but yet would generate plenty of useful data. It could form the basis of quite a bit of useful work in governance, policy, and advocacy.

Karolina Gruzel1mo

Thanks, Abeer!

MMelody ShumEndorsement1mo

Compliance and transparency are so important for governing AI. The project collects incredibly valuable data not only for accountability, but also for interoperability among EU member states, a process that requires continuous monitoring and improvement.

Karolina Gruzel1mo

Thank you, Melody! I agree with your emphasis on the importance of interoperability.

MMark AikenEndorsement1mo

Endorsed - FOIs are an important transparency and accountability mechanism. I imagine there would be broader interest from the AI safety community in viewing the FOI results, to help improve the targeting of advocacy campaigns.

Karolina Gruzel1mo

Thank you, Mark! We appreciate your endorsement.

It means a lot, especially given your extensive legal and governance expertise.

PauseAI UKEndorsement22d(edited)

Law for AI Safety have been doing excellent work to bring AI existential risks to the attention of EU politicians and catalyse discussion between the EU and China around cooperation on AI safety.

I was particularly impressed with these successes:

Raphaël Weuts, Law for AI Safety Board Member and Partnerships Lead, at a European Parliament conference on the implementation of the EU AI Act https://www.linkedin.com/posts/law-for-ai-safety_what-are-you-doing-to-advance-an-international-activity-7480969909959794688-4HZX
Raphael was also invited by the organizers of this event to help them set up coordinate with his connections working on AI at Tsinghua University (on the recommendation of an MEP).

They have a very strong team and I expect they will have a positive impact if given sufficient funding.

Anton Makiievskyi 🔸22d

Excited to fund the full ask of $20,000 @Karolina Gruzel !

Two quick questions before we distribute:

Did you receive funding from anywhere since submitting this application, or has the funding ask changed for any other reason?
Please confirm your commitment to post quarterly updates on how the project is going.

Karolina Gruzel21d

Thank you so much, Anton! This is amazing news, and we’re so excited and grateful to be able to move this project forward.

To confirm:

We have not received funding from any other sources since submitting the application, and our funding request remains unchanged.
We’re happy to provide quarterly updates on the project’s progress

MMark Aiken14d

Congratulations Karolina and the team. Fully funded now :)