Impact AI
Evaluating the societal impact of AI systems for sustainability and the public interest
So-called ‘AI for Good’ applications hold the promise of fostering both sustainability and the public interest. They are, for instance, used to monitor biodiversity, simplify the tracking of CO₂ emissions and help counter disinformation. Others are being tested for cleaning the seabed of waste. At the same time, many AI systems entail challenges. These range from the high energy and resource consumption to deficits in the transparency and accountability of automated decisions. Moreover, the AI supply chain relies on the often precarious labour of data workers. Even well-intentioned projects can have negative side effects. This raises two essential questions: How sustainable and socially beneficial are such ‘AI for Good’ projects really? And how can their broader societal value – or harm – be measured with precision?
To address these questions, the transdisciplinary research project Impact AI analyses artificial intelligence use cases from different domains that aim to contribute to sustainability and the public interest. To this end, the project has developed the Impact AI Method. This assessment systematically evaluates the actual implications of these AI systems and makes them transparent and measurable. The complete method is freely available, enabling both trained experts and organisations to make direct use of it. In doing so, the project supports the reflective and sustainable deployment of AI systems. It also provides an evidence base for informed decisions about the use of sustainable and public-interest-oriented technologies.
Research design
Our research combines scientific rigour with practical testing in a multi-stage approach. Its aim is not only to make the intended impact of ‘AI for Good’ applications measurable, but also to identify actionable pathways for their sustainable and public-interest-oriented use. It also reveals their potential, strengths and weaknesses.
Develop the Impact AI Method at HIIG, a holistic approach to measuring the impact of AI systems already in operation
Explore the method’s applicability beyond academic practice in cooperation with the partner organisations Greenpeace Germany and Economy for the Common Good (ECOnGOOD)
Examine around three use cases in depth as case studies each year
Provide participating organisations with independent feedback and recommendations for improvement
Refine the method with each case study as part of an iterative development process
Measuring impact on people, society and the planet
The Impact AI Method offers a holistic yet practice-oriented assessment of AI use cases that are already in operation (not research projects or prototypes). It focuses on their contribution to sustainable and just development. Find out how its seven steps and 25 criteria work, what participating ‘AI for Good’ projects gain and how to get involved.
The approach combines several normative principles. These include sustainability in its three dimensions, the public interest, human rights and democratic values. Despite this broad scope, it is designed to work in practice. External third parties conduct each assessment or are at least involved in it. The full method is freely available online.
Focus on democracy
Our first three case studies explore whether and how organisations can use AI to support democratic processes. They look at efforts to promote equality, support public discourse, improve access to information and make government spending more transparent. The first report, on Full Fact AI, will be published in late October or early November 2026.
Call for applications
In summer/autumn 2027, the Impact AI Fellowship brings together people from academia, journalism, civil society, public institutions and companies. Fellows learn our assessment method and apply it to a concrete AI use case. Their interdisciplinary research generates evidence on how AI applications affect society and the environment. It feeds into the debate on sustainability and the public interest.
Apply for an assessment in 2027
We welcome ‘AI for Good’ use cases from non-profit and for-profit organisations, regardless of domain or model. They should aim to serve society or sustainability and make this claim publicly. They must already be in operation (not research projects or prototypes) and remain in use for at least another year. Participants receive feedback and recommendations from an independent, non-commercial perspective.
Britta BomhauerStudent assistant: Societal Values, Transformation and AI
Laura State, PhDPostdoctoral researcher: Societal Values, Transformation and AI
Lena WinterResearcher: Societal Values, Transformation and AI
Theresa Züger, Dr.Head of Research: Societal Values, Transformation and AI
-
Sadiem YoussefFormer student assistant: Societal Values, Transformation and AI
Journal articles and conference proceedings
Fahimi, M., State, L., & Kasirzadeh, A. (2025). From Explaining to Diagnosing: A Justice-Oriented Framework of Explainable AI for Bias Detection. Proceedings of the Eighth AAAI/ACM Conference on AI, Ethics, and Society (AIES-25), 8(1), 879-892. DOI: https://doi.org/10.1609/aies.v8i1.36597 Publication details
Edited works
Ashgari, H. & Züger, T. (Eds.) (2025). Proceedings of the 2nd Workshop on Public Interest AI. Würzburg, Germany: CEUR-Workshop Proceedings. Publication details
Book contributions and chapters
Theresa Züger (2025). Wie kann KI im Sinne des Gemeinwohls eingesetzt werden? In C. Albrecht (Ed.), Künstliche Intelligenz in diakonischen Unternehmen Von Science Fiction zur Alltagsrealität, Diakonie reflektiert (pp. 14-33). Berlin, Germany: Evangelische Arbeitsstelle für missionarische Kirchenentwicklung und diakonische Profilbildung (midi). Publication details
Working paper
Mayer, K., Knaus, J., Züger, T., Fichtner, U. A., Glinka, K., Hase, J., Heller, L., Kaffee, L.-A., Kaufmann, S., Koth, S., Kowalk, D., Lipp, I., Meyer, K., Ritter, P., Waag, A.-S., & Skupien, S. (2025). Yes, we are open?! Künstliche Intelligenz verantwortungsbewusst gestalten. Weizenbaum Discussion Paper(51). DOI: https://doi.org/10.34669/WI.DP/51 Publication details
Other publications
Youssef, S., Züger, T., & Winter, L. (2026). Can AI strengthen democracy? A data collection on AI projects aiming to serve democratic processes. Digital society blog. Publication details
Winter, L., & Züger, T. (2025). Blind spot sustainability: Making AI’s environmental impact measurable. Digital society blog. Publication details
Laura State (2025). KI zwischen Klimawandel und Klimaschutz. Digital society blog. Publication details
Lectures and presentations (11)
Theresa Züger
Symposium: Demokratische KI. Berliner Wege zu Verantwortung, Steuerung und Souveränität. Technologiestiftung Berlin. Publix, Berlin, Germany: 12.05.2026 Further informationTheresa Züger
Challenges of transdisciplinary research about AI in the public interestDagstuhl Seminar 26021 - AI for Social Good. Leibniz Center for Informatics. Schloss Dagstuhl, Wadern, Germany: 05.01.2026 Further information
Theresa Züger
10. Umweltbeobachtungskonferenz. Bundesamt für Umwelt (BAFU), Umweltrat EOBC. Eventfabrik Bern Service AG, Bern, Switzerland: 28.11.2025 Further informationLaura State
Eighth AAAI/ACM Conference on AI, Ethics, and Society. Association for the Advancement of Artificial Intelligence, Association for Computing Machinery. IE University Tower, Madrid, Spain: 22.10.2025 Further informationLaura State
Wie kann KI dem Gemeinwohl dienen?DB KI-Lunch. Deutsche Bahn AG. Online, Berlin, Germany: 17.10.2025
Theresa Züger
Witten-Aachen-Workshop on AI and Sustainability (WAWAIS). RWTH Aachen, Witten/Herdecke University. Online, Aachen, Germany: 29.09.2025 Further informationTheresa Züger
EnviroInfo 2025: Open Data, Open Science and Open Society for a Sustainable Future. Gesellschaft für Informativ e.V.. University of Potsdam, Potsdam, Germany: 18.09.2025 Further informationLaura State
Greenpeace TechCamp - Sustainable Digital Transformation. Greenpeace e.V.. Online, Hamburg, Germany: 16.09.2025Theresa Züger
TACTIC Retreat / Summer School 2025. Otto von Guericke University. Otto von Guericke Universität Magdeburg, Merseburg, Germany: 20.08.2025 Further informationLaura State
Goethe-Institut. Goethe-Institut e.V.. Amerikahaus, Munich, Germany: 26.05.2025Theresa Züger
Panels (8)
Theresa Züger
KI: Macht, Mythen, Missverständnisse. Eine Gesprächsreihe über künstliche Intelligenz. SUPERRR Lab. Schiller-Bibilothek Wedding, Berlin, Germany: 23.04.2026 Further informationTheresa Züger
Künstliche Intelligenz und der neue FaschismusBerliner Landeszentrale für politische Bildung, Weizenbaum-Institut. Berliner Landeszentrale für politische Bildung, Hardenbergstraße 22–24, 10623 Berlin, Berlin, Germany: 04.02.2026 Further information
Theresa Züger
10. Umweltbeobachtungskonferenz. Bundesamt für Umwelt (BAFU), Umweltrat EOBC. Eventfabrik Bern Service AG, Bern, Switzerland: 28.11.2025 Further informationLaura State
Falling Walls Science Summit. Falling Walls Foundation. Falling Walls Science House, Berlin, Germany: 06.11.2025 Further informationTheresa Züger
Chancengleichheit jetzt! – Gemeinsam. Strategisch. Zukunftsweisend.. Bundesministerium für Arbeit und Soziales (BMAS), Bundesagentur für Arbeit. Bolle Festsäle, Berlin, Germany: 05.11.2025 Further informationTheresa Züger
re:publica 25: Generation XYZ. republica GmbH. STATION, Berlin, Germany: 27.05.2025 Further informationTheresa Züger
Leipziger Buchmesse. Forum Mensch und KI. Leipziger Messegelände, Leipzig, Germany: 30.03.2025 Further informationTheresa Züger
Moderation of workshops and panels (2)
transform_D Summit 2026. Deutsche Stiftung für Engagement und Ehrenamt. Alten Münze in Berlin, Berlin, Germany: 18.06.2026 Further information
Theresa Züger
Tech for Good Adaption Guidere:publica Berlin 2026. republica GmbH. STATION Berlin, Berlin, Germany: 19.05.2026 Further information
Theresa Züger
Organisation of events (5)
Falling Walls Science Summit. 06.11.2025. Falling Walls Science House, Berlin, Germany (International) Further information
Theresa Züger
Meaningful participation for AI in the public interestCo-producing Knowledge Under Constraint: Re-thinking Participation in Research Practice. 31.10.2025. EUniwell, Cologne, Germany. Co-Organised by: University of Cologne, Global South Studies Center Cologne (National) Further information
Theresa Züger
Omnipresent & hidden: Impact, costs, and futures of AIA workshop on the impact, costs and futures of artificial intelligence. 29.10.2025. Humboldt Institute for Internet and Society, Berlin, Germany (National) Further information
Theresa Züger, Laura State, Lena Winter
Digitaler Salon: Watt für'n Prompt29.10.2025. Humboldt Institute for Internet and Society, Berlin, Germany (National) Further information
Laura State, Katharina Mosene, Lena Winter, Judith Beyer, Sarah Spitz, Daniel Pothmann
From Rejection to Regulation: Mapping the Landscape of AI ResistanceFrom Rejection to Regulation: Mapping the Landscape of AI Resistance. 21.07.2025. Humboldt Institute for Internet and Society, Berlin, Germany (National) Further information
Theresa Züger, Can Şimşek
Nominations (1)
Theresa Züger
- 21.06. NEW ZEALAND LISTENER
The AI empire is changing life in NZ but how can we keep it — and the mega data centres it needs — under control?
- 28.03. BERLINER MORGENPOST
Berlin forscht: Was leistet Künstliche Intelligenz für das Gemeinwohl?
- 18.06.25 RESET – DIGITAL FOR GOOD
Public Interest AI: Künstliche Intelligenz für das Gemeinwohl braucht ein anderes Technologieverständnis
- 15.05.25 DEUTSCHLANDFUNK: KI VERSTEHEN
KI und Gemeinwohl: Künstliche Intelligenz für gesellschaftlichen Zusammenhalt
Project consortium
The Alexander von Humboldt Institute for Internet and Society (HIIG) conducts the transdisciplinary research project Impact AI together with its partners Greenpeace Germany and Economy for the Common Good.
HIIG coordinates the project and provides its scientific leadership. In consultation with the partner organisations, it develops the evaluation methodology. Both partners contribute their expertise in sustainability and public-interest evaluation as well as hands-on experience in both fields. Greenpeace Germany has comprehensive knowledge of environmental and climate issues and of appraising the environmental impact of technologies. Economy for the Common Good brings its understanding of sustainable economic models and its track record in public-interest-oriented business evaluation.
Funding
| Duration: | 2025 to 2029 |
| Funding: | VolkswagenStiftung as part of the Change! Fellowship programme |
Cover picture: David Clode via unsplash
Theresa Züger, Dr.
Societal Values, Transformation and Artificial Intelligence
Our research critically analyses socio-technical transformation processes. It centres on the question of how digital technologies can be more closely aligned with societal goals – particularly with regard to sustainability, the public interest and democratic values.
Impact AI: Project Council
This expert council supports the Impact AI research project in assessing public-interest AI use cases.
Call for Applications: Impact AI Fellowship 2027
Our programme offers participants a unique opportunity to explore the actual impacts of AI applications on society and the environment.
Call for Projects: Participate in our Impact AI Assessment
We are seeking organisations (NGOs, companies or research institutions) that run AI projects and would like to participate in a voluntary impact assessment.