
Photo credit: Pexels
By Kashan Ajmeri | Last updated: 19 November 2026
⏰ Applications are OPEN NOW. The call reopened 19 November 2026 at 8:00 am CET and CLOSES 20 November 2026 at 11:59 pm CET. HIIG paused the call on 30 October 2026 after receiving far more applications than expected, and the call will not stay open until 15 January 2027 as first planned. The final deadline is 20 November 2026. Have your documents ready, because the form is open for about two days only. Check the official HIIG call page, and see the open opportunities linked at the end of this post.
Applications are OPEN NOW: the call reopened on 19 November 2026 at 8:00 am CET and closes on 20 November 2026 at 11:59 pm CET. HIIG paused the call on 30 October 2026 because it received far more applications than expected, and it says the call will not stay open until 15 January 2027 as originally planned. If you want in, apply now — the form is open for about two days only. The fellowship runs for ten weeks, from 16 August to 22 October 2027, with two compulsory weeks in Berlin and the rest possibly remote. The stipend is a fixed €5,000 for the ten weeks, plus support for travel and accommodation for the in-person weeks where capacity allows. Everything below was checked against the official HIIG call on 19 November 2026; where the call is silent, this guide says so rather than guessing.
About the fellowship and the Impact AI Method
Impact AI is a transdisciplinary research project at HIIG. Its team has developed the Impact AI Method, a structured way to assess the societal and environmental impacts of an AI system in use — not the pitch deck version, and not a laboratory benchmark, but the thing as it operates in the world, with real users, real energy use, and real side effects. The project works in partnership with Greenpeace Germany and Economy for the Common Good (ECOnGOOD).
Fellows learn that method in workshops, then apply it end to end to one case: preparing and conducting interviews, analysing data, and writing a final assessment report, which they lead-author with support from the Impact AI research team. You do not need to arrive with a project of your own. HIIG runs a call for cases to match fellows with suitable AI projects, though you may propose one if you have it. That matters, because it lowers the barrier for people who have the judgement and the access — a journalist inside a newsroom, a civil servant inside a ministry — but not a ready-made research case.
One honest caveat, from the call itself: this is a research fellowship, not a coding residency and not a policy internship. The output is an evidence-based assessment report, and the method is described by HIIG as complex and demanding rigour. If you want to ship a product, this is the wrong room. If you want to produce the document that says whether a product deserves its claims, it is exactly the right one.
Who can apply
HIIG welcomes applicants from academia, journalism, civil society organisations, public institutions, and companies working on AI, sustainability, or the public interest. The call explicitly seeks people with diverse academic, professional, and personal backgrounds, and people embedded in an organisation where they could keep using the method after the fellowship ends.
A master's degree is generally required, but the call is unusually plain about the exception: not having one does not automatically disqualify you. Candidates who can show experience equivalent to a master's degree, with research experience or professional expertise in AI, sustainability, or the public interest, are encouraged to apply. The flip side is also stated plainly: applicants without experience and adequate training are less likely to be selected. An academic track record is not mandatory, but a deep interest in research is, along with the ability to work independently and to a high standard of reliability.
There is no nationality restriction stated in the official call, and no English-test score requirement is listed. Do not assume either point from third-party listings — the call itself lists neither. What it does require, without ambiguity, is attendance: participation in all in-person sessions and regular online check-ins is mandatory, and fellows sign a fellowship contract committing to the activities and outputs.
Fellowship details at a glance
| Item | Detail (official call, checked 19 Nov 2026) |
|---|---|
| Host | Alexander von Humboldt Institute for Internet and Society (HIIG), Berlin, Germany |
| Partners | Greenpeace Germany; Economy for the Common Good (ECOnGOOD) |
| Dates | 16 August – 22 October 2027 (10 weeks); mandatory online preparatory session 11 May 2027 |
| Format | Two full-time in-person weeks in Berlin (weeks 1 and 7); other weeks part-time and remote where agreed |
| Stipend | Fixed €5,000 per fellow for the ten weeks (includes any tax/social security costs in the fellow's home country) |
| Other support | Travel and accommodation for in-person participation supported individually, as capacity allows; fellows cover health insurance, visa, meals and other expenses |
| Output | Complete impact assessment of one AI project; lead authorship of a case report |
| Deadline | 19 November 2026, 8:00 am CET — call reopened (open now). 20 November 2026, 11:59 pm CET — FINAL deadline. The call was paused on 30 October 2026 and will not remain open until 15 January 2027 as first planned. |
On the stipend, be realistic. €5,000 for ten weeks is a contribution, not a salary, and the call says so in effect by listing what fellows must still cover themselves. For someone already employed, on leave, or based in Berlin, it may be workable. For someone who would need to give up paid work for ten weeks and fund their own travel, it may not be — and HIIG's note that additional costs are considered only "in exceptional cases" should be read before you apply, not after you are selected. The value here is the training, the supervision, the lead-authored report, and the network, not the money.
What fellows actually do, week by week
The programme structure, as published by HIIG, runs as follows. The preparatory phase starts three months early, because interview preparation needs lead time — a detail that tells you how seriously the interview work is taken.
- 11 May 2027 — mandatory online preparatory session. The fellowship proper has not started, but the work has.
- Week 1, 16–20 August 2027 — in person in Berlin. Workshops on the Impact AI Method and interview preparation.
- Weeks 2–6, 23 August – 1 October 2027 — independent research phase. Data collection, analysis, and assessment, with weekly check-ins and online workshops. Remote work is possible in these weeks.
- Week 7, 4–8 October 2027 — in person in Berlin. Presentation and discussion of each fellow's assessment results.
- Weeks 8–10, 11–22 October 2027 — finalising. Completing the assessment and writing the impact assessment report.
Alongside the research, fellows get comprehensive training in the method, workshops led by researchers and practitioners, individual supervision from the Impact AI research team, and an interdisciplinary peer-learning environment. HIIG also notes opportunities for future collaboration and publication based on the case-study data. None of that is guaranteed beyond the fellowship itself — treat it as a door, not a promise.
How to apply
Applications are made through the online application form on the official HIIG call page.
Note: the form is open now. It reopened on 19 November 2026 at 8:00 am CET and closes on 20 November 2026 at 11:59 pm CET — apply inside that window only.
Alongside the form, applicants submit a single combined file containing:
- CV (maximum 4 pages);
- Letter of motivation (maximum 2 pages), covering your level of AI expertise, your organisational background, and your organisation's motivation if applicable;
- If applicable, a description of the AI project you propose to assess (maximum 2 pages).
Applying costs nothing. The call lists no application fee, and the form is HIIG's own free online application — nobody will ask you to pay to apply.
Keep the motivation letter concrete. The call tells you what it is looking for: research experience or professional expertise in AI, sustainability, or the public interest; a setting where the method would keep being used; and the ability to apply a demanding method with rigour. A letter that names the kind of AI project you would want to assess — a hiring tool, an energy-optimisation system, a content-moderation pipeline — and says honestly what access and evidence you could bring will read far stronger than a general statement about caring about responsible AI. Everyone says that. Few can say what they would actually examine.
HIIG also ran two online information sessions, with no registration required: 10 November 2026 at 10:00 CET and 17 November 2026 at 15:00 CET. Both sessions have now taken place. If you missed them, the official call page still carries the full details.
Important dates
- 10 November 2026, 10:00 CET — online information session (no registration required).
- 17 November 2026, 15:00 CET — online information session (no registration required).
- 19 November 2026, 8:00 am CET— call reopened (open now).
- 20 November 2026, 11:59 pm CET — FINAL deadline, form closes.
- 30 October 2026— call paused by HIIG because of the number of applications received. Applications already sent in will still be reviewed.
- 10 November 2026, 10:00 CET and 17 November 2026, 15:00 CET— online information sessions (no registration needed). The call reopened for a short period after these sessions.
- 15 January 2027— the original deadline. It no longer applies: HIIG says the call will not remain open until this date.
- 11 May 2027 — mandatory online preparatory session.
- 16 August – 22 October 2027 — fellowship period; in-person weeks 16–20 August and 4–8 October in Berlin.
Do not plan around the old 15 January 2027 date — the call reopened only briefly. Prepare your documents in advance, and submit now, before the form closes on 20 November 2026 at 11:59 pm CET.
Official links
Apply on Official Website — Impact AI Fellowship 2027 call for applications (HIIG)
That page carries the application form, the full call text, and the information-session links. If any detail in this guide and the official page ever disagree, the official page wins — check it again before you submit.
Conclusion
The Impact AI Fellowship is a small, specific opportunity: ten weeks, one case, one report, and a method you can take back to your own organisation. Its strength is also its filter. It will suit a researcher, journalist, civil-society analyst, or public-sector professional who already works near AI and sustainability and wants the tools to assess claims properly, with supervision and a published case report at the end. It will not suit someone looking for a general AI course, a coding placement, or a fully funded ten-week salary. If you are in the first group, the €5,000 stipend and the two Berlin weeks are a fair exchange for work that genuinely sharpens how AI gets judged. The call is paused right now and reopens for about two days only, 19 to 20 November 2026, so do not plan around the old 15 January 2027 date — it no longer applies. Go to the 17 November information call with one real project in mind that you would want to assess, and be ready to apply the moment the form opens.
If you want a paid, full-time AI governance summer instead, see the GovAI Summer Fellowship 2027 in London, which pays £14,000 for three months in London, with applications closing on 22 November 2026.
Related reading: 9 AI Side Hustles for Students in 2026 · 10 Best AI Tools for Freelancers · 7 Ways Students Can Earn Money Using AI
Sources: All fellowship facts verified from the official HIIG call for applications, "Call for Applications: Impact AI Fellowship 2027" (published 9 October 2026, updated 28 October 2026), checked 29 October 2026: hiig.de/en/impact-ai-call-for-fellows.
Similar OPEN Opportunities
These fellowships and internships on AI Tools Hub are still open right now:
- GovAI Summer Fellowship 2027 in London — £14,000 salary for three months in London, 7 June – 27 August 2027. Applications close 22 November 2026 at 23:59 UK time. https://aitoolswave.blogspot.com/2026/10/govai-summer-fellowship-2027-in-london.html
- Ai2 Research Internship 2027 in Seattle — research internship with the Allen Institute for AI, 12 weeks on site in Seattle, pay of $94,320 (undergraduate), $106,110 (masters) or $140,000 (PhD) as annual rates. Applications close 8 January 2027 at 11:00pm Pacific Time. https://aitoolswave.blogspot.com/2026/10/ai2-research-internship-2027-in-seattle.html
- University of Exeter Machine Learning Ocean Forecasting Job — Postdoctoral Research Associate/Fellow in machine learning ocean forecasting at the University of Exeter, Penryn Campus, Cornwall. Salary £35,302–£44,352 per year. Applications close 10 December 2026 at 23:59 UK time. https://aitoolswave.blogspot.com/2026/11/university-of-exeter-machine-learning.html
- LawAI Summer Fellowship (US) 2027 in Washington, DC — $2,000 per week plus a $10,000 relocation stipend for ten weeks in Washington, DC, 14 June – 20 August 2027. Applications close 29 January 2027 at 11:59 pm ET. https://aitoolswave.blogspot.com/2026/11/lawai-summer-fellowship-us-2027-in.html
- Wake Forest AI Research Internship 2027 — paid 10-week summer research internship in biomedical engineering and informatics at Wake Forest University School of Medicine, 25 May – 30 July 2027, with free on-campus housing. Applications close 29 January 2027. https://aitoolswave.blogspot.com/2026/11/wake-forest-ai-research-internship-2027.html
- Radboud University NeuroAI Postdoc 2026 — postdoctoral position in NeuroAI at Radboud University, Nijmegen, building brain-inspired neural networks for smart implants. Salary €3,706–€5,760 gross per month. Applications close 1 December 2026. https://aitoolswave.blogspot.com/2026/11/radboud-university-neuroai-postdoc-2026.html
- Mozilla Foundation Fellowship 2027–2028 — up to $125,000 for AI builders, 12 months remote, Letters of Intent close 17 December 2026 at 11:59 PM Pacific. https://aitoolswave.blogspot.com/2026/11/mozilla-foundation-fellowship-20272028.html
- Microsoft Research Internship 2027: Generative Modeling at MSR Redmond — research internship in generative modeling (diffusion, image autoregressive and language models) at Microsoft Research Redmond, 12 weeks on-site, $5,690–$13,480 per month. Apply by 1 January 2027 for the fullest consideration. https://aitoolswave.blogspot.com/2026/11/microsoft-research-internship-2027.html
- University of Hong Kong Postdoctoral Fellow in AI (Medical Imaging) — Post-doctoral Fellow in artificial intelligence for medical imaging at the University of Hong Kong, Department of Electrical and Computer Engineering. Applications close 31 December 2026 (Hong Kong time), or earlier if the post is filled. https://aitoolswave.blogspot.com/2026/11/university-of-hong-kong-postdoctoral.html
- Wallenberg-NTU Presidential Postdoctoral Fellowship 2027 — up to SGD 80,000 salary plus SGD 100,000 research grant, two years split between NTU Singapore and a WASP partner institution in Sweden. Applications close 31 December 2026. Wallenberg-NTU Presidential Postdoctoral Fellowship 2027 (up to SGD 80,000 salary + SGD 100,000 grant, deadline 31 December 2026)
- IRB Barcelona AI Software Engineer Job 2026 — permanent AI Software Engineer role in cancer genomics at IRB Barcelona, building LLM and agentic AI systems. Applications close 4 December 2026. IRB Barcelona AI Software Engineer Job 2026: Permanent Role in Cancer Genomics, Deadline 4 December 2026
- Imperial Schmidt AI in Science Global Faculty Fellowship 2027https://aitoolswave.blogspot.com/2026/11/imperial-schmidt-ai-in-science-global.html
- General Motors AI Research Internship 2027 — $11,800 a month for Masters students in Sunnyvale, California, hybrid on-site 3 days a week, summer 2027. No fixed deadline, rolling applications. https://aitoolswave.blogspot.com/2026/11/general-motors-ai-research-internship.html
Only these thirteen open AI fellowship/internship posts exist on this blog right now, so all thirteen are listed here. Check the deadline on each post before you apply — deadlines change, and the official page always wins.