Skip to content
Paid starting teams wanted in three new cities
Safe AI Netherlands

Your AI Safety career starts here.

AI Safety expertise has never been more important than today. SAIN provides the community, courses and resources to help students and professionals join the AI Safety field. We provide a clear path through education, community and practical projects, then connect strong contributors to the organisations, programmes and roles where they can take the next step.

Join the community (opens in a new tab)
We are hiring paid starting teams for new chapters in Eindhoven, Delft and The Hague. Apply by 16 November. See the roles
An accelerating AI trajectory starts at the origin. Two people learn and practise while a third points up at the turning point, where the curve forks into a steep orange path and a gentler grey one.

Why SAIN exists

Too few people know how to work on AI safety, in research, in policy, or in public debate. SAIN trains more of them, connects them, and gets them started.

A free, volunteer-run path from curiosity to contribution, through courses, events, research and local chapters.

Why SAIN exists

Participate in the community

Discussion groups, hackathons, talks, and the evenings after. People meet friends and collaborators here, and often find their next step in AI Safety.

Join the community (opens in a new tab)

Our Discord community currently focuses on our SAIN Amsterdam, Utrecht and Groningen chapters. Eindhoven, Delft and The Hague are next: we are hiring their starting teams.

SAIN Amsterdam graduates celebrating outdoors with balloons and flowers
A SAIN Groningen closing evening on a rooftop terrace at sunset
A SAIN lecture filling a university auditorium
SAIN Amsterdam members around a café table with SAIN mugs
Technical AI Safety graduates beside the SAIN Amsterdam banner
An outdoor community gathering in a circle
A SAIN group session indoors

Start with a free course

Pick a track. Every programme is free and taught in person. What you join, and how it runs, depends on the chapter.

You can apply now in Amsterdam or Groningen.

See all courses

The SAIN Research Hub

Bring your academic expertise to AI Safety. Connect with researchers, develop a focused project, and take your work further with support from SAIN.

Visit the Research hub
Supervised matching
Connect with PhD+ supervisors for structured, mentored AI Safety research.
Compute & support
Access compute and logistical support for your project or open collaboration.
National network
Work with researchers, advisors, and practitioners across SAIN’s Dutch chapters.
Publication track
Build your research track record in a community with work at NeurIPS, ICLR, and other leading venues.

  1. Choose a research project

    Join a project or bring your own question. Put your academic expertise to work on AI Safety, with the people and resources to take it further.

    Read the research handbook

  2. Work with support

    SAIN helps coordinate research and arrange compute according to project needs. Develop your work with supervisor feedback or independent collaborators.

    Explore our supervisors’ research agendas

    Become a supervisor (opens in a new tab)

  3. Publish your findings

    Turn your findings into a paper, policy brief, or research tool. Get feedback on drafts, with potential support for conference travel.

    NeurIPS 2025 workshop · Spotlight

    The Anatomy of Alignment: Decomposing Preference Optimization by Steering Sparse Features

    A transparent view of preference optimisation reveals how models can learn to favour style and formatting over honesty.

    Read paper

    ICLR 2025 · Building Trust Workshop

    Self-Ablating Transformers: More Interpretability, Less Sparsity

    Training small language models to select their active components produces more specialised, interpretable circuits without sacrificing language modelling performance.

    Read paper

    NeurIPS 2025 · Regulatable ML Workshop

    EU-Agent-Bench: Measuring Illegal Behavior of LLM Agents Under EU Law

    A benchmark tests whether AI agents take unlawful actions under EU law, even when the user’s request is benign.

    Read paper
    View all publications

Apply any time. Determine your commitment per project. Remote participation is welcome.

Visit the Research hub

Build an AI Safety career

Most people who end up working on AI Safety did not plan for it. We shorten that path with mentorship, funding advice, and introductions to the labs, institutes and ministries hiring in Europe right now.

  • Technical research
  • Governance and policy
  • Field building
  • Security and compute

Internship · Existential Risk Observatory

Without this community I almost certainly wouldn’t be where I am.

Stefano Zuffi

Mapping research on AI alignment techniques and government interventions

The next decade is being decided now

You do not need a background in AI to be useful here. Every SAIN programme is free, run by volunteers, and takes you from curious to contributing, so pick one and begin.

Start with a free courseVolunteer