Rainer Mühlhoff

Rainer Mühlhoff (born 1982) is a German philosopher, mathematician and full professor for ethics of artificial intelligence at Osnabrück University, Germany.

Career

Rainer Mühlhoff studied mathematics, theoretical physics and computer science at the Universities of Heidelberg, Münster and Leipzig. His thesis was on “Higher Spinfields on Curved Spacetimes”. Afterwards he continued his studies in philosophy, gender studies and German literature in Berlin. He earned a PhD in philosophy from Free University of Berlin in 2016 with a dissertation on affect theory after Spinoza and Foucault. His work, which focused on post-structuralism, continental philosophy and critical social theory, was supervised by Jan Slaby and Martin Saar. He then worked as a research fellow at the Free University of Berlin and at the Technical University of Berlin. In 2021 he became full professor for ethics of artificial intelligence at Osnabrück University. He holds the first professorship focused on the interdisciplinary field of “Ethics of Artificial Intelligence” in Germany. In 2022, Mühlhoff was co-author of the data protection impact assessment of the German COVID-19 app “Corona-Warn-App”, which contributed in sparking a debate on the necessity and security of contact-tracing apps in Germany.

Research Focus

Mühlhoff’s research focuses on critical philosophy of digital media and social philosophy as well as ethics and critique of the digital society, big data and artificial intelligence. He has published works on privacy and data protection, intersectionality, and anti-discrimination in the context of digital technology and has an interdisciplinary approach. His philosophical approach is greatly influenced by post-structuralism and connects technology, power and subjectivity.

Positions

Human-Aided AI

Mühlhoff views machine learning systems as sociotechnical systems. According to his position, the commercial application of machine learning is structurally dependent on human participation. Unpaid labor by both users of AI systems and click workers are used by tech companies. When training AI models, media systems of human-computer interaction are deliberately designed in such a way that users produce data, like with CAPTCHAs. Current commercial AI systems would therefore not replace human intelligence but rather “capture” it.

Predictive Privacy

Mühlhoff is known for his works on data privacy in the context of artificial intelligence. In his works, he points at the societal consequences of “predictive analysis”, meaning the usage of machine learning models for the prediction of personal or unknown information about individuals: “One of the main concerns related to automated decision making based on PA [predictive analytics] is the potential contribution of this technology to stabilizing or even increasing social and economic inequalities and power differentials within societies“. According to Mühlhoff, privacy is endangered by this kind of AI applications, because predictive analytics can estimate personal information about random individuals, even some that the individual may not know themselves (e.g. disease prediction). To allow a public debate about the topic, the known value of privacy should be expanded to include “predictive privacy”. Mühlhoff suggests to include predicted or estimated information as a possible violation of privacy and not only information that an individual has willingly given somewhere. It would be irrelevant for a “predictive violation of privacy” whether the predicted information is correct, as long as the predictions are used to treat individuals differently or to make automated decisions, for example. This would differentiate the concept of predictive privacy from neighboring concepts like “group privacy”. Mühlhoff asserts that data privacy needs to be thought collectively, since the possibilities of predictive analytics are based on the sharing of data by a sufficient number of users in their day-to-day usage of digital products. This would mean that one’s own data influences others through predictive analytics.

Digital Counter-Enlightenment

In his works on user experience (UX) design, Mühlhoff diagnoses that the current evolution of digital interfaces takes away control and knowledge from the users and goes contrary to enlightened ideas of freedom and self-determination of the individual. According to Mühlhoff this happens for example through a trend of ‘sealed surfaces’ in interface design that shields away the inner workings of digital devices even from those users who want to know more.

Awards

• 2022: “Hans-Mühlenhoff-Preis” for excellent teaching of the University of Osnabrück

Publications

• Die Macht der Daten. Warum Künstliche Intelligenz eine Frage der Ethik ist. V&R unipress, Universitätsverlag Osnabrück, 2023, doi:10.14220/9783737015523.
• Predictive Privacy: Collective Data Protection in the Context of AI and Big Data. In: Big Data & Society, 2023, doi:10.1177/20539517231166886, S. 1–14.
• Mühlhoff, Rainer, und Theresa Willem. Social Media Advertising for Clinical Studies: Ethical and Data Protection Implications of Online Targeting. In: Big Data & Society, 2023, doi:10.1177/20539517231156127, S. 1–15.
• Mühlhoff, Rainer, und Hannah Ruschemeier. Predictive Analytics und DSGVO: Ethische und rechtliche Implikationen. In: Telemedicus – Recht der Informationsgesellschaft, Tagungsband zur Sommerkonferenz, 2022, S. 38–67.
• Prädiktive Privatheit: Kollektiver Datenschutz im Kontext von Big Data und KI. In: Künstliche Intelligenz, Demokratie und Privatheit, 2022, doi:10.5771/9783748913344-31, S. 31–58.
• Predictive Privacy: Towards an Applied Ethics of Data Analytics. In: Ethics and Information Technology. 23 (2021), doi:10.1007/s10676-021-09606-x, S. 675–690.
• Automatisierte Ungleichheit: Ethik der Künstlichen Intelligenz in der biopolitische Wende des Digitalen Kapitalismus. In: Deutsche Zeitschrift für Philosophie. 68 (2020), doi:10.1515/dzph-2020-0059, S. 867–890.
• Human-Aided Artificial Intelligence: Or, How to Run Large Computations in Human Brains? In: New Media & Society. 22 (2020), S. 1868–84. doi:10.1177/1461444819885334.
• Menschengestützte Künstliche Intelligenz. Über die soziotechnischen Voraussetzungen von Deep Learning. In: ZfM – Zeitschrift für Medienwissenschaft 21(2) (2019), S. 56–64. doi:10.25969/mediarep/12633.
• Mühlhoff, Rainer, Anja Breljak, und Jan Slaby, Hrsg. Affekt Macht Netz. Auf dem Weg zu einer Sozialtheorie der digitalen Gesellschaft. Bielefeld: transcript, 2019, ISBN 978-3-8376-4439-5.
• Immersive Macht. Affekttheorie nach Foucault und Spinoza. Frankfurt am Main: Campus, 2018, ISBN 978-3-593-50834-4.
• Digitale Entmündigung und ‚User Experience Design‘. In: Leviathan – Berliner Zeitschrift für Sozialwissenschaft. 46 (2018), doi:10.5771/0340-0425-2018-4-551, S. 551–74.