Skip to content

We are hiring

See open positions
Safe AI Netherlands

Project Hub

The SAIN Project Hub handbook

How research through SAIN works, from first application to published output. It sets out what researchers, supervisors, and the Research Team can expect from each other.

Version 1.0 · 29 April 2026 · research@safeainetherlands.org

Written by Alexander Müller (Director), Thomas Brcic (Former SAIN Groningen Co-Director), and Ilija Lichkovski (Research Lead).

What the Project Hub is for

SAIN exists to raise awareness of the full spectrum of existing and potential harms from AI, contribute to shaping mitigation priorities through ongoing discourse, and support the realisation of effective solutions. The Project Hub is where that turns into research: a platform where supervisors, collaborators, and students connect to advance AI safety work on a global and interdisciplinary scale.

This handbook applies to researchers and students working under the hub, supervisors affiliated with SAIN, collaborators and visiting researchers on open projects, and anyone outside SAIN who wants to understand how the hub runs.

Almost any topic counts, if it reduces AI-related risk

We welcome any project that meaningfully contributes to AI safety: technical, governance, or conceptual. If you are unsure whether a topic counts, the default answer is yes, provided you can articulate a plausible pathway from your work to reducing AI-related risk. We mean risk broadly: risk to lives, to equality, to rights. To make the scope concrete, the hub especially encourages work in the following areas, without being limited to them.

Technical alignment and interpretability

  • Mechanistic interpretability
  • Representation learning and feature sparsity for safer control
  • Steering and controllability of LLMs and other foundation models
  • Adversarial robustness and red-teaming of models
  • Detection and mitigation of deceptive or misaligned behaviours
  • Scalable oversight, debate, constitutional AI, and related methods
  • Robustness of RL agents and multi-agent systems
  • Safety benchmarks and evaluation metrics
  • Agent foundations
  • Neuroscience-inspired alignment

Societal impacts, governance, and policy

  • AI governance, regulation, and standards, with emphasis on the EU and Dutch context
  • AI and democratic processes: misinformation, polarisation, information integrity
  • AI and labour: job displacement, economic and social impacts
  • Privacy, surveillance, and data protection in AI deployment
  • Risk assessment for high-stakes domains such as CBRN, cybersecurity, and critical infrastructure

Meta and foundational topics

  • Research methodology and evaluation in AI safety
  • Benchmarking AI systems' persistence, persuasion, or autonomy
  • Forecasting and scenario analysis for transformative AI
  • Epistemics, information hazards, and responsible communication in AI safety

Four principles hold on every project

  1. Intellectual honesty. Always communicate accurately what is known, unknown, and uncertain.
  2. Proactively consider near-term and long-term harms, as well as dual-use concerns: ways the same work could both protect and cause harm. Section 4 covers this in depth.
  3. Maintain a supportive environment across disciplines and seniority levels. There is zero room for harmful authority.
  4. Feedback, critique, and review are essential tools, not personal attacks. When receiving feedback, remember it is aimed at the work. When giving it, attack the ideas, not the person.

Be extra wary of these principles when deadlines approach. Everyone is fallible, and when compromising on a principle is the difference between publishing and not publishing, holding the principle matters more than the publication.

Before you apply

There are two ways to do research through SAIN

The hub currently runs in two modes. Everything later in this handbook applies to both unless it says otherwise.

Supervised research matching.

Students and early-career researchers apply for a guided project with mentorship from one of SAIN’s supervisors. SAIN provides logistical and financial support, such as compute, and projects aim to evolve into publishable research. Supervisors are experienced researchers with an interest or track record in AI safety, usually at PhD level or more senior (we write this as PhD+ below), though exceptions are possible. We welcome supervisors from any discipline.

Open collaboration.

Not all research needs formal supervision. SAIN connects researchers directly and provides support, such as compute, so you can focus on the work. If you have a research proposal, submit it via the form on our website. If you want to join an existing project, reach out to that project’s contact person. Nobody plays the supervisor role on these projects, but one person leads, usually whoever designed the project.

Anyone with sufficient background is eligible

  • Both modes are open to anyone with sufficient background: Bachelor's, Master's, and PhD students, researchers from industry, and people with no formal affiliation.
  • Remote participation works.
  • To become a supervisor, we mostly ask that you are PhD+, though in exceptional cases this can differ.

You can apply at any time

  • The hub works on a rolling basis. Apply whenever you are ready; if there is an opportunity and you are a good fit, you will be accepted.
  • For supervised research matching, we get back to you within a working week.
  • For starting an open collaboration, we appreciate a fully fleshed-out proposal before you contact us, but if necessary we can help shape it.

From application to running project

Onboarding starts with one form

Every new researcher gets a rough overview of current projects, groups, and supervisors, an explanation of this handbook and where it lives, and, if relevant, access to compute or other support. Join our Discord by filling in the onboarding form; the invite and the information above follow from there.

Fill in the onboarding form (opens in a new tab)

Open collaborations begin with a research proposal

Before a project is approved as a Project Hub open collaboration, the researcher completes a research proposal. The full structure below is only needed when the goal is a full paper and collaboration; for other outputs, drop what you think is unnecessary. A mechanistic interpretability tool, for example, does not need a title and abstract.

  1. Title, problem statement, and motivation
  2. Background and related work
  3. Research questions and hypotheses
  4. Proposed methodology
  5. Safety and risk considerations: dual-use, misuse, and infohazards, meaning information that could cause harm if widely shared
  6. Needed resources: potential datasets, compute estimates, and how many collaborators for how much time
  7. Timeline and milestones
  8. Expected outputs

SAIN’s Research Team reviews the proposal and either approves it as a Project Hub project, with or without revisions, or declines it.

Supervised matching begins with an expression of interest

For supervised research matching, or to join an existing open collaboration, fill in the expression of interest form or email research@safeainetherlands.org.

Register your interest (opens in a new tab)

What happens once you are approved

  • You are brought into contact with the relevant people. For supervised matching, a supervisor is formally assigned and a project channel is created in the Discord. For joining an existing open project, you are officially added to the project and its Discord channel. For a new open project, we add it to our website, add you to the Discord, and promote the project so others can join.
  • Expectations around time commitment (for example 5 to 10 hours per week versus near full-time) and meeting cadence are set. We encourage communication through the Discord, which you join via the onboarding form; other channels can be agreed if preferred.
  • The project is registered internally and kept up to date as it progresses.
  • Where applicable, the Research Team sets up a GitHub repository under the SAIN organisation and gives project members access.

During the project

Researchers drive the work; supervisors steer it

Three roles keep a project moving, and each owes the others something concrete.

Researchers and students are expected to

  • Take primary responsibility for the day-to-day progress of the project.
  • Show up to supervisor meetings prepared, with a clear agenda, results, or specific questions.
  • Fill in the weekly check-in form.
  • Communicate early about blockers or personal constraints.
  • Follow SAIN’s core principles (section 1).

Supervisors are expected to

  • Provide regular guidance and mentorship, and high-level project direction, stepping in where needed.
  • Help ensure the project is well-scoped and realistic.
  • Meet at the agreed frequency, usually weekly or biweekly, and show up prepared.
  • Give timely feedback on drafts and research directions.
  • Watch out for projects stuck in unproductive directions, and for researcher wellbeing issues such as stress or burnout.
  • Raise concerns with the Research Lead or team where appropriate.

The Research Team is responsible for

  • Maintaining up-to-date documentation, including this handbook.
  • Onboarding new researchers and supervisors.
  • Keeping track of Project Hub projects and outputs.
  • Acting as the contact point, via research@safeainetherlands.org or the Discord.
  • Project management, unblocking, and research support sit here too; direct your questions to the Research Team first.

The rhythm is weekly

During a project, the standard pattern is:

  • Weekly or biweekly meetings: discuss recent progress and blockers, refine direction and scope, review experiments, results, and drafts.
  • Weekly check-ins: a short form noting what was tried, progress, and next steps. It keeps accountability light and lets the team see how projects are going.
  • Contact with the Research Team whenever necessary, via research@safeainetherlands.org or the Discord.

Communicate early, precisely, and kindly

  • Use the agreed channels, ideally the Discord, for project communication.
  • Respond within a reasonable timeframe: 2 to 3 working days for non-urgent matters.
  • Meetings are the supervised researcher's responsibility: have a simple agenda, start with a brief update since last time (the weekly check-in helps you keep track), and end with concrete next steps.
  • Be explicit about availability: travel, family time, exam periods.
  • Default to kind, precise, and honest communication.

Agree milestones up front

Each project should specify intermediate milestones. As an example, not a fixed rule:

  • Weeks 1 to 3: getting to know each other and the project, literature review.
  • Weeks 4 to 8: first experimental proof of concept.
  • Weeks 8 to 10: internal discussion of initial results and what to improve.
  • Weeks 10 to 12: implementing improvements, new results.
  • Weeks 12 to 14: writing up the final report and submission.

SAIN supports the work but does not pay for it

This applies mostly to technical projects.

  • SAIN does not provide stipends or salaries. Participation is voluntary and educational.
  • If funding for conference trips, workshop registrations, or travel is needed, we may be able to help within reason.
  • Once a project is specified, the Research Team strives to give it easy access to compute and other resources. How much compute SAIN can help with depends on the project and is agreed at the start.
  • We encourage computationally inexpensive projects wherever possible.

Integrity, AI tools, and dual-use

Integrity

  • Cite sources properly, after making sure you know what they actually say, and avoid plagiarism.
  • Be honest about experimental results, limitations, and negative findings.
  • Keep enough documentation that others could roughly reproduce your work.

Responsible use of AI tools

We are not against using AI tools such as LLMs for coding or writing. Where a researcher has considered it carefully and it clearly helps, we encourage it. However:

  • Always check outputs carefully. Do not uncritically trust generated content.
  • Avoid feeding sensitive or confidential data into tools unless you are confident they handle data responsibly.
  • Where relevant, disclose your use of AI tools.

Dual-use and infohazards

Work on AI safety can have paradoxical effects: it can make AI systems, the AI ecosystem, or people inside it less safe. These effects can play out in complex ways that are impossible to fully foresee. At minimum, spend some time thinking about whether your work could significantly increase capabilities or misuse risk if widely shared.

  • If in doubt, discuss it with your supervisor and/or SAIN's Research Lead before public dissemination.
  • Possible mitigations: redacting sensitive details, internal-only or restricted-access reports, or delaying publication until the risks are better understood.

Understanding matters more than publishing

The Project Hub should be a place where personal wellbeing is taken seriously and support is there for anyone who needs it. AI safety research, like any research, can be intellectually and emotionally demanding, especially under the publish-or-perish pressure that is unfortunately far too common.

  • Be honest about workload and stress levels with your supervisor, and supervisors with their researchers.
  • Take regular breaks and keep boundaries between work and rest.
  • Remember that what matters most is gaining a better understanding of AI safety, not whether one publishes.
  • Speak to the SAIN team if conflicts or issues arise.

Finishing and publishing

A project is complete when the question is answered

  • The main research question has been adequately addressed, including negative or null results, and the work has been published somewhere in a presentable form.
  • If relevant, infohazard and dual-use considerations have been treated appropriately.

Final deliverables can take several forms

Each project specifies one or more final deliverables. For instance:

Authorship reflects contribution

  • Authorship should reflect substantial intellectual and implementation contributions.
  • Author order is handled internally within the project team.
  • SAIN and the Project Hub should be acknowledged in publications.
  • SAIN may promote your work as enabled by the Project Hub.

After the project

  • Output and other relevant resources, such as data, code, notes, and drafts, are stored in SAIN’s project archive.
  • Write a short note on what went well and what could have gone better, ideally shared with the SAIN community.
  • The project may be extended as a follow-up.

If something goes wrong

We hope you never need this section, but it matters that the path is written down before anyone needs it.

When to notify someone

  • A supervisor or researcher relationship breaks down, for example through persistent unavailability or unresolved conflict.
  • There is suspected misconduct, such as data falsification, harassment including sexual harassment, or plagiarism.
  • There are serious concerns about whether the project's direction is still aligned with AI safety.
  • Project scope, expectations, or time commitments have become misaligned.

The path runs from direct to leadership

  1. Researcher and supervisor: try to resolve it directly first.
  2. SAIN's Research Lead, or anyone in the Research Team.
  3. SAIN's Co-Directors, if the conflict involves the Research Lead or the Research Team.

Possible remedies and sanctions

Depending on the severity and nature of the issue, possible actions include:

  • Clarifying expectations and adjusting milestones.
  • Reassigning the supervisor or researcher to another project.
  • Re-scoping or pausing the project.
  • Removal from the Project Hub, or from SAIN roles, in severe cases.

If you supervise

Supervisor expectations are in section 4. The administrative procedure for a supervised project is:

  1. The research proposal is prepared as described in section 3.
  2. Researchers are assigned to supervisors and projects.
  3. The supervisor sends the research proposal, together with the list of researchers on the project, to research@safeainetherlands.org. Include contact details, and each researcher’s Discord username and GitHub username where applicable, as a minimum.
  4. If applicable, a repository is created in the SAIN GitHub organisation, and researchers and supervisors get collaborator access.
  5. Every Sunday, researchers fill in the weekly check-in form; every second Sunday, the supervisor fills in the bi-weekly check-in form. Both forms are linked under Deadlines in section 4.

Interested in supervising AI safety research through SAIN? We mostly ask that you are PhD+, exceptions possible, and we welcome any discipline.

The hub is open on a rolling basis.

Apply whenever you are ready. If there is an opportunity and you are a good fit, you will be accepted.

Register your interest (opens in a new tab)Become a supervisor (opens in a new tab)