Meet the Fellows

Meet the new Friedrich Schiedel Fellows who work on building new bridges between social sciences, technology, and other disciplines. Their interdisciplinary research projects, under the motto "Human-Centered Innovation for Technology in Society" focus on how technologies can be developed responsibly, human-centered, and democratically, serving the public good.

Auxane Boch: Psychology Impact Assessment for Interactional Systems: Defining the Evaluation Scope (PSAIS)

This research project aims to address the lack of frameworks for systematically assessing the diverse psychological impacts of AI. By adopting a participatory approach and considering cultural values, the project seeks to develop a multi-cultural mapping framework for evaluating the psychological impact of AI systems. The research will involve workshops and consultations with stakeholders from various sectors and regions to define evaluation criteria. The project will contribute to the development of concrete recommendations for action by providing a culturally-informed framework that can guide the responsible development and application of AI technologies. The impact of the project extends to academic disciplines, partner institutions, societal stakeholder groups, and policy actors. It will foster interdisciplinary knowledge exchange, stimulate discussions on standardization, and contribute to the reduction of potential inequities arising from technology adoption. The project aims to be a change agent by actively contributing to the implementation of the recommendations and ensuring user well-being and trust in AI systems.

Efe Bozkir: Echoes of Privacy: Exploring User Privacy Decision-Making Processes towards Large Language Model-based Agents in Immersive Realities

User privacy concerns and preferences have been researched extensively in the context of various technologies, such as smart speakers, IoT devices, and augmented reality glasses, to facilitate better privacy decision-making and human-centered solutions. With the emergence of generative artificial intelligence (AI), large language models (LLMs) have started being integrated into our daily routines, where models are tuned with vast amounts of data, including sensitive information. The possibility of embedding these models in immersive settings brings a plethora of questions from privacy and usability point of view. In this project, through several user studies, including crowdsourcing ones, we will explore privacy concerns and preferences towards LLM-powered and speech-based chat agents for immersive settings and inference likelihood of alarming user attributes. The findings will help understand the privacy implications of such settings, design informed consent procedures that support users in immersive spaces that include LLMs, and facilitate privacy-aware technical solutions.

Baris C. Cantürk: Future Finance Law Hub (“F2L_Hub”)

The Future Finance Law Hub (“F2L_Hub”) is a project aimed at becoming a policy-maker hub at the intersection of IT law and commercial law, established at TUM, with the primary area of influencing Germany and the European Union in the medium to long term. Central to its modus operandi is the aggregation of prominent stakeholders from academia and industry across multidisciplinary domains, including law, finance, and IT, to generate significant outputs.
As delineated in the foregoing summary, F2L_Hub embodies a mission of considerable magnitude, delineating medium and long-term objectives aimed at institutionalizing a tradition in this field. Thus, this fellowship is poised to function as the catalyzing force behind the establishment of F2L_Hub, leveraging both financial resources and access to an excellent academic environment. Through this esteemed fellowship, a noteworthy organizational milestone will be attained, facilitating the procurement of requisite funding and partnerships vital for the execution of ancillary processes.

Daryna Dementieva: Harmful Speech Proactive Moderation

Offensive speech remains a pervasive issue despite ongoing efforts, as underscored by recent EU regulations aimed at mitigating digital violence. Existing approaches primarily rely on binary solutions, such as outright blocking or banning, yet fail to address the complex nature of hate speech. In this work, we want to advocate for a more comprehensive approach that aims to assess and classify offensive speech within several new categories: (i) hate speech that can be prevented from publishing by recommending a detoxified version; (ii) hate speech that necessitated counter speech initiatives to persuade the speaker; (iii) hate speech that should be indeed blocked or banned, and (iv) instances mandating further human intervention.

Mennatullah Hendawy: Setting up the Future with Sustainable Choices: GenAI Support in Resolving Multi- Stakeholder Conflicts in Sustainable Critical Metals & Minerals Development

This project outlines an innovative approach to resolve multi-stakeholder conflicts in the sustainable development of critical metals and minerals essential for decarbonization efforts. Recognizing the complexities and sustainability challenges within the supply chains of these materials, particularly those sourced from the Global South / emerging economies, the project proposes a digital platform leveraging reactive machine AI (RM-AI) and generative AI (Gen-AI) with human-in-the-loop functionalities. This platform is designed to facilitate transparent and inclusive discussions among public/community representatives, government, and industry stakeholders, ensuring a balanced consideration of environmental, economic, and social sustainability targets. Through co-developing a concept for an interactive, game- based decision-making tool powered by Gen-AI, the project aims to identify common interests, model sustainability trade-offs, and find consensus solutions that align with the societal goals of reducing inequality and promoting economic growth with decent work conditions. The project's integration of RM and Gen AI aims to bridge the gap between technical and non-technical decision-makers, enhancing stakeholder engagement and trust in AI-driven processes, thereby aligning closely with the fellowship’s mission of human- centered innovation and interdisciplinary collaboration for the public good.

Franziska M. Poszler: Research-based theater: An innovative method for communicating and co-shaping AI ethics research & development

This project will implement a creative approach to conducting, educating, and communicating AI ethics research through the lens of the arts (i.e., research-based theater). The core idea revolves around conducting qualitative interviews and user studies on the impact of AI systems on human ethical decision-making. It focuses specifically on exploring the potential opportunities and risks of employing these systems as aids for ethical decision-making, along with their broader societal impacts and recommended system requirements. Generated scientific findings will be translated into a theater script and (immersive) performance. This performance seeks to effectively educate civil society on up-to-date research in an engaging manner and facilitate joint discussions (e.g., on necessary and preferred system requirements or restrictions). The insights from these discussions, in turn, are intended to inform the scientific community, thereby facilitating a human-centered development and use of AI systems as moral dialogue partners or advisors. Overall, this project should serve as a proof of concept for innovative teaching, science communication and co-design in AI ethics research, laying the groundwork for similar projects in the future.

More information on the project can be found here:https://www.ieai.sot.tum.de/research/moralplai/

Malte Toetzke: Developing the Google Maps for the Climate Transition

I envision to develop the Google Maps for the climate transition. Business leaders and policy makers need more comprehensive and timely evidence to accelerate industrial development of climate-tech effectively. With recent advances in Artificial Intelligence (AI), it is now possible to develop models that generate such evidence at large scale and in near real-time. In the project, I will analyze the global network of organizations collaborating on climate-tech innovation. The network is based on processing the social media posts of organizations via large language models (LLMs). It includes key public and private actors and spans various types of climate technologies (e.g., solar, hydrogen, electric vehicles) and types of collaborations (R&D collaborations, demonstration projects, equity investments). I will use the fellowship to conduct in-depth analyses generating valuable insights for managers and policy-makers on facilitating innovation clusters. Furthermore, I plan to operationalize the information retrieval and processing enabling analyses in real-time.

Chiara Ullstein: Participatory Auditing (in cooperation with Audit.EU)

The EU AI Act posits that providers (management and developers) of high-risk AI systems have to undergo conformity assessment. The conformity assessment encompasses several measures that are supposed to corroborate that a system is legally compliant, technically robust, and ethically sound, and can be considered ‘trustworthy AI’. The project ‘participatory auditing’ aims to contribute to the project Audit.EU (1) by exploring how companies can leverage their learnings from established compliance practices such as for the GDPR and (2) by proposing participation as an approach to source AI Act compliance-relevant information from suitable stakeholders to increase inclusivity and mitigate risks of discrimination. Participation is considered to enhance the process of achieving compliance through a comprehensive testing and feedback process. Based on learnings from established compliance measures, a framework for performing auditing in a participatory manner and in accordance with the EU AI Act will be developed and evaluated. The primary goal of the framework is to serve developer teams as a guideline.

Niklas Wais: Law & AI: Navigating the Intersection

Most areas of law that should in principle be relevant for AI currently leave many intersectional questions unanswered. The reason for these open questions is that jurisprudence cannot pursue its task of incorporating AI into the existing dogmatics because it lacks sufficient technological understanding. At the same time, developers lack knowledge of the law and therefore only base their design decisions on performance, but not compliance with e.g. data protection or anti-discrimination law. Although students from various professional backgrounds want to learn more about the underlying interface issues, truly interdisciplinary educational material is missing. My project will address this and transform the rare specialist expertise that currently only exists at TUM into a freely available online course. By fostering interdisciplinary collaboration between law and technology and sharing cutting-edge knowledge as effectively as possible, the project seeks to promote the responsible use of AI for the benefit of society.

The research project "Using AI to Increase Resilience against Toxicity in Online Entertainment (ToxicAInment)", by Prof. Dr. Yannis Theocharis (Chair of Digital Governance), funded by the Bavarian Research Institute for Digital Transformation (BIDT), explores the spread of extremist, conspiratorial and misleading content on social media, investigating how this content is embedded through entertaining content. It aims to deepen the understanding of the impact of this content on user behavior by combining entertainment theories, visual communication and toxic language with AI methods. This project makes an important contribution to analyzing and combating online toxicity. More information can be found on the project page or in the BIDT press release.

A two-day workshop bringing together experts in the field

Content moderation and free speech in the digital realm - and how to balance them - are key topics for researchers, philosophers, public officials, NGOs, and, of course, social media platforms and users. At the TUM Think Tank, we had the pleasure of hosting a number of international experts in this field. The group came together for two full days focused on analyzing this pressing issue, exchanging ideas, and presenting empirical research from the perspectives of governance, industry, and political behavior.

From ideological biases in content moderation and the politics of platform regulation to citizens’ preferences on how online harmful speech can be curved and regulated, and the efficacy of labeling content as AI-generated, the workshop covered a wide range of topics, stressing the need for a transnational conversation about content moderation.

Panel discussion

In a thought-provoking panel together with Benjamin Brake (Federal Ministry of Digital Affairs and Transport), Friedrich Enders (TikTok Germany), Andreas Frank (Bavarian Ministry of Justice), and Ruth Appel (Stanford University), we discussed the complexities of defining harmful speech and taking action against it, how platforms are audited and how they balance transparency with user privacy and free expression when it comes to content moderation decisions.

The conversation centered on the division of responsibility for content moderation and the transparency of enforcement from the key stakeholders involved. It was noted that while the German government is responsible for smaller platforms not covered under the Digital Services Act (DSA), the European Commission is responsible for larger ones like X or TikTok.

• While the precise ways in which tech companies should deal with harmful speech based on the definitions and guidelines provided by the DSA are clouded by some vagueness, a common theme in the discussion was the necessity for transparency in content moderation decisions and the need to always take context into consideration. Based on the conversation, vagueness in defining harmful speech can be seen as a flexible way of dealing with it by tech companies and governments. Researchers, on the other hand, pointed out that it can also be problematic, especially when it comes to its precise detection through automated methods.
• In addition, Friedrich Enders shed light on TikTok's content moderation process. The platform uses a combination of AI and human review to quickly remove harmful content. Conscious of the fact that some harmful, e.g. graphic content, may still be in the public interest, such content may remain on the platform for documentary, educational, and counter-speech purposes, but would be ineligible for recommendation to users on TikTok’s For You feed.
• The panel also highlighted the challenge of balancing freedom of expression, user privacy, and user safety, with TikTok stressing their commitment to both principles, with the government strongly advising that the importance of upholding freedom of expression is such that one should always opt for freedom of speech when one is doubtful about how borderlines cases need to be moderated.

The Chair of Digital Governance co-jointly organized the workshop at the Munich School of Politics and Public Policy, the University of Oxford, the Technical University of Munich, and the Reboot Social Media Lab at the TUM Think Tank.