About us
We investigate the ethical and philosophical aspects concerning the development of a certification for AI applications
Conferences and Talks
Access the archive of our conference presentations and find out the latest news about our upcoming talks and workshops
Publications
Check out our publications. Find out more about our research focus and get the latest updates on our findings and activities
Latest News
Publication of paper on Portuguese LLM
Dr. Nicholas Kluge Corrêa is excited to announce his latest research together with our colleagues Aniket Sen, Sophia Falk and Shiza Fatimah (all University of Bonn), which
advances open-source Portuguese NLP. This study tackles the challenge of building strong language models for low-resource languages by introducing GigaVerbo, a large, high-quality dataset that enables Portuguese language modeling at a scale surpassing 500B tokens. The team developed the Tucano model series, which outperforms comparable multilingual models and even exceeds Llama-3.2 on several benchmarks. The paper also highlights limitations in current evaluation benchmarks, demonstrating that many widely used metrics don't accurately reflect model scaling in Portuguese. All datasets, models, code, and logs are openly available to support a fairer and more sustainable future in Portuguese NLP.
Nicholas Kluge Corrêa invited by OECD as an expert
Dr. Nicholas Kluge Corrêa was recently invited by the OECD to participate in the OECD-African Union (AU) Artificial Intelligence (AI) Dialogue 2.0 – AI in Africa: From strategy to implementation. At this high-impact event in Egypt, Dr. Corrêa will be working on the Developing an African Charter on Trustworthy AI workshop, alongside delegates of the OECD-African Union and other invited experts. This interactive session focusses on collaboratively shaping the contours of a charter to guide AI governance in Africa, adapting global principles to the African context to ensure cultural and contextual relevance.
Latest Work
Want to know more about the ongoing work in our project?
Take a look at our "Catalogue of General Ethical Requirements for AI Certification", which is available on ArXiv. Given the large number of attempts to operationalise ethical principles (for AI), we decided to take an applied approach and present tools that developers can use to design more ethical software. The tools are an attempt to integrate values into the design of AI applications. We present ethical considerations on fairness, privacy and data protection, security and robustness, sustainability, transparency and explainability and truthfulness. The catalogue will soon be available as a white paper in English and German.
Workshop 2024 "Leave no one behind"
On July 11th and July 12th we are hosting our workshop "Leave no one behind: who is affected by current AI development, governance, standardization and regulation processes?" Day one will take place at the CST, Day two at the IZPh. Online participation is possible. Registration is now open.
14th Bonner Wissenschaftsnacht May 16 & 17, 2024
At the Bonner Wissenschaftsnacht, we look forward to discussing with you the relationship between freedom, 'Freigeist', and privacy. Julia Maria Mönig will present the results of the TRA 4 project "Rethinking Privacy after this Pandemic". Find us in tent 6 at the 'Zeltstadt des Wissens' (Münsterplatz).
New Team Member
We are happy to welcome our new team member Nicholas Kluge Corrêa. Nicholas has a master's degree in engineering and and is currently finishing his PhD in philosophy jointly at the University of Bonn and the Pontifical Catholic University of Rio Grande do Sul. Welcome to the team, Nicholas!
Zertifizierte KI on tour
After attending the meeting of the AG Digitale Ethik of the Initiative D21 with an Input by Prof. Dr. Janina Loh, the philosophy team went on a cruise on the Rhine on the occassion of the Digital Ethics Summit 2023. Our flagship project "Zertifizierte KI" was presented there by our colleague Rebekka Görge, a data scientist at Fraunhofer IAIS.
Publication of the Manifesto for the Future of Privacy
We are launching the Manifesto on the occasion of the World Health Assembly, which will begin on May 21, 2023, in order to highlight that next to important health policy decisions the protection of privacy and other ethical aspects needs to be guaranteed.
New Publication: "Standardizing fairness-evaluation procedures: interdisciplinary insights on machine learning algorithms in creditworthiness assessments for small personal loans"
New publication in the Springer Journal AI and Ethics.
AI, Human Values, and Meaningful Human Control - Registration now open!
The conference will take place in Bonn on June 22-23, 2023.
Call for Abstracts: AI, Human Values, and Meaningful Human Control
The conference aims at addressing ethical issues of the impact of machines on human choices and value formation. Submission deadline on by Feb. 20th, 2023.
Launch of the Second Edition of the German Standardization Roadmap for Artificial Intelligence
Dr. Sergio Genovesi and Dr. Julia Maria Mönig contributed to the second edition of the German Standardization Roadmap for Artificial Intelligence that was published on 9 December 2022. The Roadmap for Artificial Intelligence provides an essential basis for the further development of artificial intelligence.
