Artificial intelligence is growing faster than most rules can keep up with. Who controls powerful AI systems. Who is responsible when they go wrong. How can normal people and governments still have real power over this technology. These are huge questions. The KU Leuven AI Safety Lab wants smart researchers to help answer them.
This Lab is now opening its Visiting Fellowship 2027. Only five fellows will be picked. If your work touches AI safety, AI governance or the risks of advanced AI, this is worth a close look.
At a Glance
| Detail | Information |
|---|---|
| Host | KU Leuven AI Safety Lab |
| Location | Leuven Belgium |
| Duration | Three months |
| First cohort | March to May 2027 |
| Monthly allowance | €2000 |
| Number of fellowships | Five |
| Format | Fully in person |
| Application deadline | 31 October 2026 |
What This Fellowship Actually Is
This is a three month research fellowship based in Leuven, Belgium. The first group of fellows will stay from March to May 2027. During this time, you work on your own research project. You are not alone though. You join a full team at the Lab and work inside their daily research life.
The fellowship is fully in person. Remote work is not an option.
Who This Fellowship Is Built For
This is not only for computer scientists. The Lab wants people from many fields. These include:
- Philosophy
- Law
- Political science
- Economics
- Psychology
- Computer science
- Related fields
If your work looks at power control rules or the future of technology from any angle this fellowship may fit you.
The Two Main Research Areas
When you apply you must pick one research programme.
Theories of Victory and Viable Endgames
This area looks at the good side of advanced AI. What would a strong and safe future with AI actually look like. What steps do we need to get there?
Democratic Control over AI Companies
This area studies something very real right now. A small number of big AI companies hold a lot of power. This research area looks at legal, political and system-based ways that normal democratic societies could still hold real control over that power.
There is also a third option called Other. If your project does not fit the two named areas, you can still apply here. But your work must show a clear and strong link to AI safety.
What You Actually Get
Here is what the fellowship gives you:
- A desk at the Lab’s office in Leuven
- Access to KU Leuven libraries and facilities
- A monthly allowance of €2000 for living costs
- Formal status as a visiting scholar at KU Leuven
One important point. This is not a normal job. You do not get an employment contract. You still handle your own taxes and social security on your own. Also, you will join regular talks at the Lab. You can share your work in progress at internal seminars. You may also get the chance to present at the Lab’s yearly conference and summer school.
Who Can Apply
This fellowship is open to a wide group of people:
- PhD candidates
- Postdoctoral researchers
- Faculty members
- Researchers working outside normal academic jobs
If you work outside academia, you still need to show that this fellowship truly serves real research goals, not just general interest.
The Lab also says clearly that it wants more applications from women, minorities and other groups who are often underrepresented in this field.
One more key point. Your project must be something you can truly finish inside three months. Pick a topic that is focused, not something huge that needs years of work.
What Is Expected From Fellows
This fellowship asks for real active participation. You must work from the Leuven office regularly. This is not built for a remote-style arrangement.
By the end of your three months, you are expected to:
- Finish a full research paper
- Post it as a preprint on arXiv, PhilPapers or a similar site
- Send it to a target journal, conference or academic venue that you listed in your plan
- Present your work at the Lab
- Share your findings after the fellowship ends
- Take part in Lab activities such as the annual conference or summer school
Working with another author is allowed but not required. The Lab is happy to support team research when it makes sense for the project.
What Your Application Must Include
You need to prepare these documents:
- CV Maximum two pages. Must include your list of publications.
- Cover Letter Maximum one page. Explain why you want this fellowship, why you fit the Lab and what you hope to gain.
- Research Plan Maximum 1500 words. Your plan should cover:
- The research question and why it matters for AI safety
- Your planned method or approach
- Your theory of change, meaning the real impact you expect your work to have in academic study, policy or the wider AI safety field
- The journals, conferences or workshops you plan to submit your paper to
- Why your project can truly be done inside three months
- Which research programme you are applying under
How Selection Works
The Lab looks at several things when picking fellows:
- How strong and original your research idea is
- How well it fits one of the Lab’s research programmes
- How clear and believable your theory of change is
- Whether your plan is truly realistic for three months
- What you personally can add to the Lab’s research community
Key Dates to Remember
| Milestone | Date |
|---|---|
| Call published | 21 September 2026 |
| Application deadline | 31 October 2026 |
| Interviews | November 2026 |
| Final results | End of November 2026 |
| First cohort in Leuven | March to May 2027 |
Why This Fellowship Is Worth Taking Seriously
AI safety is a fast-growing and increasingly important field. This fellowship gives you real research time, direct supervision, access to a strong university environment and a monthly income to support you while you work. For many researchers, that combination is hard to find elsewhere.
This fellowship fits best if you already have a clear research question, a solid plan and a strong sense that you can finish a real paper within three months. It is not built for broad or unfocused ideas.
How to Make Your Application Stronger
A few simple steps can really help your chances:
- Keep your project tightly focused so it truly fits within three months
- Make your theory of change very clear and specific
- Show a strong and honest connection between your idea and your chosen research programme
- Explain clearly why KU Leuven is the right place for this exact project
- Set realistic publication targets instead of overly ambitious ones
- Make the AI safety relevance of your work obvious, not something a reader has to guess
The Bottom Line
The KU Leuven AI Safety Lab Visiting Fellowship 2027 offers a rare chance. You get real funding. You get a strong academic home. You get to work alongside people across many fields on one of the biggest questions of our time: how the world should manage powerful AI.
With only five spots open and the deadline set for 31 October 2026, start preparing your research plan now rather than waiting until the final days.