Research Areas

Humanity looks up to the scientific community to provide answers and solve both practical and conceptual challenges of contemporary society. Scientific decisions directly and deeply influence human lives. Thus, securing responsible scientific decisions becomes an imperative. The Carl Friedrich von Weizsäcker Center is interested in the investigation of the three theoretical foundations of responsible science: ethics of science, epistemology of scientific research, and the humanization of scientific work.

Ethics of science is concerned with scientific misconduct, e.g., plagiarism and fraud, with questions about boundaries of science, i.e., which research is morally permissible, and with the abuse of science for commercial purposes. Epistemology of scientific research investigates optimal ways of scientific knowledge production and is concerned with challenges such as the replication crisis, biases in science, communication patterns, etc. In particular, the Center is exploring the social dimensions of scientific knowledge acquisition. One of the topics addressed is the question of responsible allocation of epistemic resources in science.

Finally, scientific results are influenced by the treatment researchers receive from society. During politically turbulent times, but also at the time of increasing pressure and focus on productivity, there is a danger that scientists become regarded as commodities. In order to secure stimulating work conditions for research, one has to explore the treatment of scientists and comprehend what their fundamental rights and needs are.

Relevant Publications:

• Kneer, M. and Stuart, M. T.:  Playing the Blame Game with Robots. Companion of the 2021 ACM/IEEE International Conference on Human-Robot Interaction. DOI: 10.1145/3434074.3447202, 2021
• Sikimić, V., Nikitović, T., Vasić, M. & V. Subotić. Do Political Attitudes Matter for Epistemic Decisions of Scientists? Review of Philosophy and Psychology. 2020.
• Sikimić, V.: Science Policy. Column in the June issue of The Reasoner. 2020.

The new AI, based on stastic analysis of big data, has a whole series of promising successes to show (e.g. on image recognition; medicine; self-driving cars; etc.). However, AI is not a "panacea" either. It will not be expected to solve the stopping problem, for example, and if efficient factorization algorithms were within its reach, Internet banking would probably have to be stopped as well. We discuss computational issues for which modern AI cannot (yet?) be seen to provide solutions. In particular, we are interested in the question how it could be formally shown that these problems cannot be solved by AI methods.

Two workshops on this topic have already been organized at the annual conference of the Gesellschaft der Informatik: Conceptual Challenges for AI.

Relevant Publications:

• Kahle, R. and Mainzer, K. Grenzen der KI - theoretisch, praktisch, ethisch. Springer Verlag, 2022.
• Kahle, R.: Wozu (ver)führt uns die neue KI? In Anna Strasser, Wolfgang Sohst, Ralf Stapelfeldt, and Katja Stepec (Herausgeber): Künstliche Intelligenz: Die große Verheißung. Philosophische Kontexte, vol. 8, Xenomoi, 2021.
• Kahle, R.: Primzahlen als Herausforderung. In R. Reussner, A. Koziolek, and R. Heinrich (editors): INFORMATIK 2020. Lecture Notes in Informatics, Gesellschaft für Informatik, S. 719-727, 2021.
• Mainzer, K.: Wie sicher ist KI? In R. Reussner, A. Koziolek, and R. Heinrich (editors): INFORMATIK 2020. Lecture Notes in Informatics, Gesellschaft für Informatik, S. 695-718, 2021.
• Kahle, R. and Mainzer, K.: Konzeptionelle Herausforderungen für die KI. In R. Reussner, A. Koziolek, and R. Heinrich (editors): INFORMATIK 2020. Lecture Notes in Informatics, Gesellschaft für Informatik, S. 693-694, 2021.
• Mainzer, K.: Künstliche Intelligenz: „Post Corona“. 2020.
• Mainzer, K.: Grundlagen, Forschung und Philosophie: „Post Corona“. 2020.

We investigate questions concerning the foundations of mathematics, especially following the Hilbert school and the scientific work of Paul Bernays. Proof-theoretic research concerns both mathematical theories and structural questions, especially in the area of proof-theoretic semantics.

Former Collaborators: Roberta Bonacina; René Gazzari; Wilfried Keller

Relevant Publications:

• Santos, P. and Kahle, R.: k-Provability in PA. Logica Universalis, 2021. (DOI 10.1007/s11787-021-00278-1)
• Kahle, R. & Rathjen, M. (eds.): The Legacy of Kurt Schütte. Springer, 2020.
• Kahle, R.: "Sehr geehrter Herr Professor!'' Proof Theory in 1949 in a Letter from Schütte to Bernays. In R. Kahle and M. Rathjen (editors): The Legacy of Kurt Schütte, pages 3-19. Springer, 2020.

Artificial Intelligence ethics is an extremely lively research area spanning multiple ethical issues in multiple branches of technology, from the use of offensive language by chatbots to the invasive collection and analysis of data on digital platforms, from the risk of encoding biases and prejudices in data and AI to the difficulty of attributing responsibility for the harm caused by autonomous systems. At the Center, we pursue research on how ML models influence the fairness of decision-making, and how they re-shape the ethical and legal issues surrounding privacy. We also consider the use of machine learning in evaluating science and structuring research teams, for example under which conditions machine learning could and should be used as part of grant review in science. 

Relevant Publications:

• Pour un développement des IAs respecteux de la vie privée. Blog Binaire, Le Monde, 2021.
• AI Social Sciences and Humanities Website

Relevant Events and Talks:

• Definitions of Algorithms: Computer Science and Law Seminar

• Algorithmic Fairness and the Epistemic Limitations of Bureaucracy, Philosophy of Science and Methodology Colloquium, Tübingen, 2021.

• What Kind of Programs are Large Language Models? The case of GPT-3". PROGRAMme Project, Bertinoro, Italy, 2021.

• Algorithmic Grant Review: Benefits and Limitations, Philosophy of Science Meets Machine Learning, Tübingen, 2021.

• The Potential of Machine Learning in Grant Review: Predicting Project Efficiency in Physics, European Philosophy of Science Association - EPSA 2021, Turin, Italy, 2021.

Intensional phenomena are still a challenge for logical analysis. Here we try to provide more adequate tools with new approaches, among others from proof theory. The topic is also treated from the specifically linguistic perspective.

Relevant Publications:

• Kahle, R.: The Intensional Structure of Epistemic Convictions. In Loek Cleophas and Mieke Massink (editors): Software Engineering and Formal Methods. SEFM 2020, Collocated Workshops. Lecture Notes in Computer Science, vol. 12524, Springer, 2021.