.png)
4 months ago
9
Reflecting on my time in the European Commission DG CNECT A2 Artificial Intelligence Regulation and Compliance in the AI Office
This article is written from my personal perspective and does not represent the views of the European Commission. These are my opinions in a blog format to reflect on my time in the AI Office. Who am I? My name is Alex Moltzau and I have been a Policy Officer Seconded National Expert for the last year in the European AI Office. My background is in social data science and AI for public services, and I worked full-time with AI policy nationally in Norway for five years prior to my current job.
There is no secret that the start of the AI Office has been an adventure. Always when new things are started and people are brought together there is a degree of navigation and charting out new waters. It has been a lot of hard work, and I will share with you some reflections from the year that has passed.
First and foremost, I really love the AI Office and the work. If you are into AI policy and law it must be one of the best places you can be globally. There are many colleagues to learn from, dedicated staff with deep domain knowledge (legal, technical or social) and plenty of challenges to come to terms with.
The work is also highly meaningful considering the position that Europe is taking on responsible AI these days.
You will not find a boring week if you are an AI policy wonk (a person who takes an enthusiastic or excessive interest in minor details of political policy). Nerding out with colleagues on details of policy is possible because so many people care about the topic and are updated on the latest developments in policy and technology.
In my first naïve blog post after the first week in the European AI Office I thought I may be able to blog every week or keep actively engaging socially, and it turned out that I was very wrong. Oh, was I wrong…
The tempo at which the European AI Office has been going and the amount of work has been significant.
I really struggled in the start to cope with the workload, and it can still be difficult at times. The work is not for the faint of heart due to the pressures from all sides and the short deadlines.
One of my fondest memories is still from the first day in the AI Office when I started and when the office went ‘live’. It felt new and exciting. Throughout the year there have also been many memorable moments, and memorable people, worth telling you about.
I started writing the draft of this article after I had been in the AI Office for five months, it was supposed to have six months in the title, but I am now finishing it after a year. On the honest side, this is how it can turn out when there is a lot of work with constant deadlines. I also moved to Brussels with my wife, a newborn son and a two year old daughter so that has certainly been part of it.
I hope that when you are reading this it makes our work more human, and that you realise we are just people doing our best to make a better environment for responsible AI in Europe and beyond. Perhaps you may even consider joining us in this endeavour wherever you may be.
What did I actually do?
Well, a lot of work… But here are a few points to give you an idea:
- Prepared and launched the AI Pact pledges signatory process and worked with outreach. Proactive work towards the AI Act with OpenAI, Google, Anthropic, Microsoft, Airbus, IKEA, and other large or small companies. More than 100 companies were brought together in the Berlaymont for the launch of the commitments. To date (June 2025) more than 200 companies have signed the AI Pact Pillar II and more than 2000 organisations are part of the broader community. We have also established a webinar series and a range of workshops.
- Contributing to the work on AI Regulatory Sandboxes including overview and background for the implementing act. Recently started drafting the implementing act on AI regulatory sandboxes. Shared a 60+ page discussion document and 20 page background paper for the implementing act with Member States. Coordinating the AI Board subgroup on AI regulatory sandboxes. It has been a great experience in listening to the needs of the various states around Europe for policy implementation.
- I helped design the new onboarding process for new recruits at the AI Office with a buddy system and welcoming procedures. Helping with recruitment, reading through more than 100 CVs for technical experts. Helping with the interviewing of both technical experts and policy officers.
- Writing policy briefings for a range of leadership in the European Commission. Including Executive Vice President Henna Virkkunen, Director General Roberto Viola, previous Executive Vice President Margrethe Vestager and more.
- Started up and run the social committee for our unit in the AI Office. Including organising the New Year’s Party, carnival, bowling, BBQ, food hall visit, anniversary and more. Recently also helping to set up two social clubs internally. The cycling club and the coffee club.
I do not want to sugercoat it, although you will find plenty of chocolate in Brussels, it is not a sweet life every day.
Life in the European Commission is filled with many challenging moments with a range of stakeholders. Yet you share those moments with some incredible people.
The European AI Office faces constant criticism and scrutiny from many sides ranging from industry to governance. The parliament, council and leadership keeps us running. Large and small companies keep us on our toes. Civil society and states keeps us continuously accountable.
But hey, that’s democracy, which is worth celebrating and being grateful for.
It has honestly been difficult at times, so the holidays were very welcome whenever they came around, yet it has been great to work with such a dedicated team.
Life is beautiful, and of course I can complain, but at the same time I feel very lucky to get to share this journey with people in our team.
Days can be filled with wonder and curiosity, yet there is also an extensive range of administrative work that just has to be done. There is also the added layers of accountability for the work done with requests for information and policy briefs. These pop up, often in the middle of various deadlines.
1. The people in the AI Office care deeply about their work and work hard to get things right
When I moved to Brussels and talking to my local hairdressers or people in shops, many seem to have an impression that the people inside the European Commission are lazy bureaucrats. At least in the AI Office this is so far removed from the reality that it is astounding, with people putting in a generous amount of time and effort every single day. Some of the hardest working people out there, as far as I have observed.
In my past I have experienced working in startups, nonprofits, big four consulting (KPMG) and research (public sector, university based). So far, the most intense is the AI Office. Likely due to the workload and current political focus on AI from the European Commission, European Parliament and European Council combined with the extensive press coverage alongside recent geopolitics focused on AI.
There is also people with a range of extensive legal, technical, social and/or policy expertise that are some of the biggest nerds on AI policy, law and models that I have ever met. If you have a group of pure paper pushers in your mind, think again. You will find a textured cast of characters, with several leading roles that are inspiring to work with.
2. The European AI Office is truly taking shape
On the first day I became familiar with the structure of the different units in the AI Office as I came to the meeting where these were introduced.
The European AI Office is located within the European Commission in DG CNECT Directorate A.
Directorates-General (DGs) are policy departments, which are responsible for different policy areas. DGs develop, implement and manage EU policy, law, and funding programmes. In addition, service departments deal with particular administrative issues. Executive agencies manage programmes set up by the Commission.
A1 is based in Luxembourg, while the rest of the units are based in Brussels. There have been several pushes for recruitment with a range of technical experts. Several of these units were transferred or changed from existing structures with the exception of the A3 unit for AI Safety that is completely new.
It is clear that there is a collaboration in different ways between Directorate Generals (the departments of the European Commission). One example is the collaboration with DG Sante in the subgroup under the AI Board with Medical Device Regulations and AI Act overlap. There will also be a bigger focus on health going forward.
The new employees that are coming into the AI Office are, according to what I have experienced, highly competent and some can even be said to be considerable talents in the field of AI overall. In addition, a lot of talented people I know or have heard about in the field of AI policy and law joined the office over the last year.
3. The international engagement runs deep
There is a range of international engagement, the EC has a wide network of stations. With delegations ranging from San Francisco, Berlin to Japan. Also with different companies, and I got to experience this when I helped set up the AI Pact and gather the signatories together.
- There is also an international team that constantly engages with different communities around the world. The Head of International Juha Heikkilä knows a wide range of languages fluently (7 or more I think), while also having a strong background in AI and robotics as the previous Head of Unit for this effort.
- There are people from the European Commission spread around the world who give detailed in-depth policy reports. These are very useful.
- We have key people placed in some of the leading environments for AI that report back with interesting insights. They work on a personal level and bring back important intel that feeds into our work.
4. AI has become hot geopolitics with investments and discussions on high-levels
With the advent of a new year and new leadership in the US the Stargate investment claiming to activate $500 billion in funding for AI was announced. Shortly after France announced €110 billion for AI and the EU with a targeted €200 billion investment.
5. There is a strong wish to engage with the scientific community that is reflected in some of the structures in the AI Act and the AI Office
There is a range of ways that the AI Office engages with the scientific community ranging from the personal to the structural.
The scientific panel is one that we are directly working to set up through the AI Act.
“The Artificial Intelligence (AI) Act envisages the establishment of a scientific panel of independent experts to advise on, and assist the AI Office and national market surveillance authorities with, implementing and enforcing the AI Act.”
There is currently a call for experts for the scientific panel:
The Code of Practice for general-purpose AI has also engaged a broad set of stakeholders :
“The Code of Practice established by the AI Office in the EU aims to create guidelines for the use of artificial intelligence across the Union. This includes ensuring that AI information is kept up-to-date, describing the data used to train AI, and managing potential risks. The Code serves as a tool under the AI Act to bridge the interim period until harmonized European standards are adopted.”
Furthermore, there is support for startups and SMEs working on AI products and services for researchers:
As well as the AI-on-demand platform for AI and robotics:
6. We strive to listen to civil society, public, industry and government
There are of course various ways to engage with the AI Office.
I previously described some of these in a separate blog post last year, but summarised we do engage continuously with public input, policy briefs, information requests and democratic procedures. This is a very real part of our everyday life.
7. The dialogue with the European Parliament and the European Council is important for a democratic approach to AI
Concerns are raised. Challenges of the concerns that are raised in a political perspective are translated and negotiated. There are a range of dialogues between these Parliament, Commission and the Council. In this way Member States can raise their concerns both through their elected Members of European Parliament and the political leadership in each country. Additionally the AI Board has representatives from Member States on a governance level with subgroups with specific public officials involved related to specific areas of collaboration.
8. You get responsibility from day one
Writing speeches and policy briefs for political leadership, working directly on files with repercussions for the rollout of policy mechanisms on the entirety of the European area, and contributing to the overall work on one of the largest AI regulatory efforts in the world. When you come into the AI Office you are not put to peripheral tasks, according to what I have experienced, you are put to work straight away. We have to much to do for nothing to be done, every day we have to pull together to move ahead and every person in the team is needed.
9. Working towards global accountability in governing responsible AI is a worthwhile endeavor
For me, I had the joy of working directly on a file (AI Pact) with the chief negotiator from the EU towards the Council of Europe (CoE). His knowledge and experience of navigating complex relationships with respect was highly inspiring and present in another file we were working on together. Although I have not worked on the CoE, it was clear that thoughts and ideas from the AI Act process came into several discussions. We are working with the European market and citizens, yet it is clear that we have to consider global collaboration. We cannot grow into an AI continent in isolation, we have to build bridges and share pathways towards increased accountability. With the new year of 2025 and new geopolitical changes, we have to maintain a momentum to ensure responsible governance of AI that protects rights and ensures safe products.
10. Getting some of the largest technology and AI companies in the world to sign a piece of paper or generally agree is a rollercoaster
Having taken part in a process to get some of the largest technology companies in the world to sign a piece of paper together, with the work done on the AI Pact and for proactive implementation of the AI Act with the signatories, I came to realise it is a rollercoaster. As much as it is exciting to follow these companies and their developments, it is clear that they have different perspectives and priorities. Most organisations have some ideas regarding responsible AI, yet we have to look at how we can make it work in practice, and this is where the wagon rolls upwards. Seeing CEOs and Directors from some of the leading tech companies in the world gathered in one room is a rush, yet the work behind attempts at aligning is immense. For colleagues working with a 1000 stakeholders in the Code of Practice I could also see these challenges.
11. When things get tough you need a team with good spirit, competence and kindness
When you find a task difficult you can likely get help from colleagues who have done things before or clever people who will troubleshoot with you.
It is the small acts of kindness that matter. When I came into the office, sleepless from a newborn son and a two-year old daughter, having moved my life to a new country, I was not sure if I would feel at home. Would I feel welcome in this new place and this new office? During my first weeks I was invited out to lunch by colleagues and always helped with new tasks as well as systems by leadership.
“Be kind, for everyone you meet is fighting a battle you know nothing about.” — Wendy Mass
For me, having received so much from colleagues, it gives a lot to give back too. I think this is a virtuous circle. Despite the timelines and running after deliverables we have to look sideways and see how we can be there for each other.
Together across Europe we are a special kind of team. We share a common area, yet we also share the beautiful diverse range of cultural experiences and interpretations. This year has also proved to be inspiring when meeting the talented people in governments spread across different nations. We have so much to offer each other and so much to share that can enrich our lives.
12. When one door opens more doors open
You know the saying.
The phrase “when one door closes, another opens”.
It is often attributed to Alexander Graham Bell. He was also credited with patenting the first practical telephone.
This saying suggests that when one opportunity is lost, another one becomes available.
However, the full quote states:
“When one door of happiness closes, another opens; but often we look so long at the closed door that we do not see the one which has been opened for us”
This reflects the idea that while we may on the one hand focus on missed opportunities, we should on the other hand remain open to new possibilities that arise.
We need to see the door that has been opened for us.
For me, being part of the AI Office has opened new possibilities.
The door that opened has led to many new doors that I could not have imagined.
As a twelfth point I have three open doors I wanted to mention.
The first open door, meeting interesting people in the policy circles and being listened to. There were a range of meetings that were great, and too many to mention. Yet to get the chance to meet a deeply engaged Digitalisation and Governance Minister for lunch, and for another time dinner then seeing some of the changes that followed that were implemented. Of course, just sharing available information and perspectives.
The second open door is for learning in new environments.
My personal engagement is a Visiting Policy Fellowship at the University of Cambridge. This received ethical approval by my leadership, and I was really happy that they also thought it was fine to have this scientific engagement — being connected to the research community — next to my job.
The third open door is for the opportunity to become more European.
Working in the European Commission gives you the possibility of gaining friends and aquaintances across Europe and learning from their stories. Going to a few places and immersing yourself in this wonderful region we live in, and all these places to cherish.
Thank you for reading my blog post and a few of my points.
I know it got a bit long, but glad you read up to this point. If you did, feel free to reach out to me on LinkedIn and tell me what you think, or your ideas within AI policy and law.
