Open topics

Food systems are a major contributor to global environmental degradation, with dietary choices accounting for roughly 25% of greenhouse gas emissions (Crippa et al., 2021). Despite growing awareness, consumers still struggle to assess the sustainability of food products at the point of sale (Camilleri et al., 2019). This is partly due to the complexity of interpreting sustainability labels and partly due to behavioural barriers. Even when information is available, the intention-behaviour gap often persists (Carrington et al., 2014).

The rise of online grocery shopping allows consumers to make food choices from home (GroceryDoppio, 2023) and opens new opportunities to support more sustainable decisions through targeted interface design (De Bauw et al., 2022). However, real-time sustainability feedback, a potentially powerful intervention, remains underexplored in digital food environments (Shin et al., 2020).

This thesis aims to conceptualize an intervention that combines product-level sustainability information (Clark et al., 2022; Poore & Nemecek, 2018; Ritchie et al., 2022) with real-time, basket-level feedback in an online grocery setting. For example, shoppers could be shown a composite score summarising the overall environmental impact of their basket, based on criteria such as greenhouse gas emissions, land use, or animal welfare. An online experiment should be conducted to test the influence of this feedback on consumer decision-making. If needed, the intervention can be implemented as an interactive prototype or mock-up.

The experiment will be implemented in collaboration with researchers from the Nürnberg Institute for Market Decisions (NIM), who will co-supervise the thesis.

Prerequisites:

• Required:
  • Enrolled Master Student in International Information Systems, International Business Studies, Economics, Marketing, etc. Bachelor students can only be considered upon careful consideration of timeline feasibility.
  • Interest in advancing research in the food domain.
  • Willingness to write their thesis in English.
• Highly Desirable:
  • Previous experience in experimental methodology, e.g., through our courses “Experimentelle Verhaltensforschung in Data Science (EVIDS)” (Bachelor) or the seminar “Information Systems for Behavior Change (ISBC)” (Master).
  • Previous knowledge and/or experience in the application of statistical methods (e.g., ANOVA, LMM, etc.)

References

• Camilleri, A. R., Larrick, R. P., Hossain, S., & Patino-Echeverri, D. (2019). Consumers underestimate the emissions associated with food but are aided by labels. Nature Climate Change, 9(1), 53-58.
• Carrington, M. J., Neville, B. A., & Whitwell, G. J. (2014). Lost in translation: Exploring the ethical consumer intention–behavior gap. Journal of business research, 67(1), 2759-2767.
• Clark, M., Springmann, M., Rayner, M., Scarborough, P., Hill, J., Tilman, D., … & Harrington, R. A. (2022). Estimating the environmental impacts of 57,000 food products. Proceedings of the National Academy of Sciences, 119(33), e2120584119.
• Crippa, M., Solazzo, E., Guizzardi, D., Monforti-Ferrario, F., Tubiello, F. N., & Leip, A. J. N. F. (2021). Food systems are responsible for a third of global anthropogenic GHG emissions. Nature food, 2(3), 198-209.
• De Bauw, M., De La Revilla, L. S., Poppe, V., Matthys, C., & Vranken, L. (2022). Digital nudges to stimulate healthy and pro-environmental food choices in E-groceries. Appetite, 172, 105971.
• GroceryDoppio. (2023, February 7). January 2023: State of Digital Grocery Performance Scorecard. Performance Scorecard. https://www.grocerydoppio.com/performance-scorecard/january-2023-state-of-digital-grocery-performance-scorecard
• Poore, J., & Nemecek, T. (2018). Reducing food’s environmental impacts through producers and consumers. Science, 360(6392), 987-992.
• Ritchie, H., Rosado, P., & Roser, M. (2022). Environmental Impacts of Food Production. Environmental Impacts of Food Production. https://ourworldindata.org/environmental-impacts-of-food
• Shin, S., Van Dam, R. M., & Finkelstein, E. A. (2020). The Effect of Dynamic Food Labels with Real-Time Feedback on Diet Quality: Results from a Randomized Controlled Trial. Nutrients, 12(7), 2158.

Food systems are a major contributor to global environmental degradation, with dietary choices accounting for roughly 25% of greenhouse gas emissions (Crippa et al., 2021). Despite growing awareness, consumers still face difficulties in identifying more sustainable food options during the shopping process (Camilleri et al., 2019). Even when information is available, the complexity of sustainability indicators and behavioural barriers often prevent this knowledge from translating into action (Carrington et al., 2014).

Online grocery platforms offer new opportunities to support sustainable decision-making through digital tools and interface design. Among these tools, recommender systems are widely used to suggest products based on user behaviour or preferences (de Bauw et al., 2022; Jesse & Jannach, 2021). While often optimised for sales, such systems can also be designed to promote sustainable choices (Felfernig et al., 2023).

This thesis aims to conceptualize a sustainability-sensitive recommender system for online grocery environments. The system could prioritise or highlight products with lower environmental impact, based on indicators such as greenhouse gas emissions, land use, or animal welfare (Clark et al., 2022; Poore & Nemecek, 2018; Ritchie et al., 2022). Students are encouraged to explore or propose their own criteria and recommendation logic. An online experiment should be conducted to evaluate the influence of the recommender system on consumer decision-making. If needed, the system can be implemented as an interactive prototype or mock-up.

The experiment will be implemented in collaboration with researchers from the Nürnberg Institute for Market Decisions (NIM), who will co-supervise the thesis.

Prerequisites:

• Required:
  • Enrolled Master Student in International Information Systems, International Business Studies, Economics, Marketing, etc. Bachelor students can only be considered upon careful consideration of timeline feasibility.
  • Interest in advancing research in the food domain.
  • Willingness to write their thesis in English.
• Highly Desirable:
  • Previous experience in experimental methodology, e.g., through our courses “Experimentelle Verhaltensforschung in Data Science (EVIDS)” (Bachelor) or the seminar “Information Systems for Behavior Change (ISBC)” (Master).
  • Previous knowledge and/or experience in the application of statistical methods (e.g., ANOVA, LMM, etc.)

References

• Camilleri, A. R., Larrick, R. P., Hossain, S., & Patino-Echeverri, D. (2019). Consumers underestimate the emissions associated with food but are aided by labels. Nature Climate Change, 9(1), 53-58.
• Carrington, M. J., Neville, B. A., & Whitwell, G. J. (2014). Lost in translation: Exploring the ethical consumer intention–behavior gap. Journal of business research, 67(1), 2759-2767.
• Clark, M., Springmann, M., Rayner, M., Scarborough, P., Hill, J., Tilman, D., … & Harrington, R. A. (2022). Estimating the environmental impacts of 57,000 food products. Proceedings of the National Academy of Sciences, 119(33), e2120584119.
• Crippa, M., Solazzo, E., Guizzardi, D., Monforti-Ferrario, F., Tubiello, F. N., & Leip, A. J. N. F. (2021). Food systems are responsible for a third of global anthropogenic GHG emissions. Nature food, 2(3), 198-209.
• De Bauw, M., De La Revilla, L. S., Poppe, V., Matthys, C., & Vranken, L. (2022). Digital nudges to stimulate healthy and pro-environmental food choices in E-groceries. Appetite, 172, 105971.
• Felfernig, A., Wundara, M., Tran, T. N. T., Polat-Erdeniz, S., Lubos, S., El Mansi, M., … & Le, V. M. (2023). Recommender systems for sustainability: overview and research issues. Frontiers in big Data, 6, 1284511.
• Jesse, M., & Jannach, D. (2021). Digital nudging with recommender systems: Survey and future directions. Computers in Human Behavior Reports, 3, 100052. https://doi.org/10.1016/j.chbr.2020.100052
• Poore, J., & Nemecek, T. (2018). Reducing food’s environmental impacts through producers and consumers. Science, 360(6392), 987-992.
• Ritchie, H., Rosado, P., & Roser, M. (2022). Environmental Impacts of Food Production. Environmental Impacts of Food Production. https://ourworldindata.org/environmental-impacts-of-food

Generative Artificial Intelligence (GenAI) has rapidly become pervasive in our everyday lives. However, seminal studies are already showing a range of negative side-effects of GenAI use on cognitive engagement and abilities, risking cognitive atrophy and overreliance (Kosmyna et al., 2025; Lee et al., 2025; Schoeffer et al., 2025; Zhai et al., 2024). Therefore, it is imperative that human-AI interfaces are designed in a way that promotes deep engagement and the critical reflection of outputs (Yatani et al., 2024).

This thesis is meant to contribute to this endeavor by developing and experimentally testing a design intervention in a (mock-up) LLM interaction interface like ChatGPT that promotes reflection and critical thinking. Students may contribute their own specific ideas for study contexts and settings. Possible interventions could be elements that induce deliberate friction, increase transparency or provide metacognitive scaffolds.

The experiment will be implemented in collaboration with researchers from the Nürnberg Institute for Market Decisions (NIM), who will co-supervise the thesis.

Prerequisites:

• Required:
  • Enrolled Master Student in International Information Systems, International Business Studies, Economics, Marketing, etc. Bachelor students can only be considered upon careful consideration of timeline feasibility.
  • Interest in advancing research in the food domain.
  • Willingness to write their thesis in English.
• Highly Desirable:
  • Previous experience in experimental methodology, e.g., through our courses “Experimentelle Verhaltensforschung in Data Science (EVIDS)” (Bachelor) or the seminar “Information Systems for Behavior Change (ISBC)” (Master).
  • Previous knowledge and/or experience in the application of statistical methods (e.g., ANOVA, LMM, etc.)

References

• Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2506.08872

• Lee, H.-P. (Hank), Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 1–22. https://doi.org/10.1145/3706598.3713778

• Schoeffer, J., Jakubik, J., Vossing, M., Kuhl, N., & Satzger, G. (2025). AI Reliance and Decision Quality: Fundamentals, Interdependence, and the Effects of Interventions. Journal of Artificial Intelligence Research, 82, 471–501.

• Yatani, K., Sramek, Z., & Yang, C.-L. (2024). AI as Extraherics: Fostering Higher-order Thinking Skills in Human-AI Interaction (Version 2). arXiv. https://doi.org/10.48550/ARXIV.2409.09218

• Zhai, C., Wibowo, S., & Li, L. D. (2024). The effects of over-reliance on AI dialogue systems on students’ cognitive abilities: A systematic review. Smart Learning Environments, 11(1), 28. https://doi.org/10.1186/s40561-024-00316-7

Generative Artificial Intelligence (GenAI) has rapidly become pervasive in our everyday lives. However, seminal studies are already showing a range of negative side-effects of GenAI use on user decision-making, risking cognitive atrophy and overreliance (Kosmyna et al., 2025; Lee et al., 2025; Schoeffer et al., 2025; Zhai et al., 2024). However, studies have also shown that carefully designed human-AI interfaces can mitigate such effects and promote deep engagement and the critical reflection of outputs (Yatani et al., 2024).

In this highly relevant emerging research field, an overview of evidence on GenAI overreliance is necessary. The aim of this thesis is to conduct a literature review to investigate which factors can increase or decrease the occurrence of overreliance in human-(Gen)AI interactions. Students may contribute their own ideas to adjusting the scope.

Prerequisites:

Enrolled student at WiSo. If your study program is not a part of WiSo, please check your program requirements whether our team is allowed to supervise your thesis.

References:

• Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2506.08872
• Lee, H.-P. (Hank), Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 1–22. https://doi.org/10.1145/3706598.3713778
• Schoeffer, J., Jakubik, J., Vossing, M., Kuhl, N., & Satzger, G. (2025). AI Reliance and Decision Quality: Fundamentals, Interdependence, and the Effects of Interventions. Journal of Artificial Intelligence Research, 82, 471–501.
• Zhai, C., Wibowo, S., & Li, L. D. (2024). The effects of over-reliance on AI dialogue systems on students’ cognitive abilities: A systematic review. Smart Learning Environments, 11(1), 28. https://doi.org/10.1186/s40561-024-00316-7

Applications are closed for summer semester 2025.BackgroundAccording to the extended mind hypothesis, human cognition extends beyond the brain and nervous system to the body and environmental tools (Clark & Chalmers, 1998). Using technological tools to facilitate cognitive processes is often referred to as “cognitive offloading.” Or “distributed cognition” (Risko & Gilbert, 2016). In the age of generative AI and its increasing capabilities, the possibilities for humans to offload energy- and time-consuming cognitive processes to AI are numerous and evolving rapidly.

Research Gap

In many human-AI collaboration scenarios, humans remain the “final authority” to accept or reject AI recommendations. Therefore, using AI for task completion not only necessitates controlling and monitoring one’s own cognitive processes (often referred to as metacognition) but also evaluating AI’s processes and outputs (Dunn et al., 2021; Tankelevitch et al., 2024). Despite the increasing metacognitive demands of generative AI and the relevance of accurate metacognition to prevent under-/overreliance on AI outputs, there is a lack of research investigating the role of human metacognition in (successful) Human-AI collaborations.

Thesis Goals

Bachelor’s and Master’s theses on this topic aim to investigate the dual role of metacognition: first, in the decision-making process of when to offload cognitive tasks to AI; and second, in how humans evaluate AI outputs once collaboration occurs. Research should explore how targeted interventions can be designed to enhance metacognitive accuracy during human-AI collaboration. The goal is to gain conceptual and empirical insights that will advance our understanding of human-AI cognitive partnerships and inform future design approaches.

Research Approaches

• (Systematic) Literature Review
  • Conduct a (systematic) literature search exploring interventions (such as cognitive forcing strategies) that support metacognitive monitoring/control and prevent excessive cognitive offloading.
  • Review interdisciplinary research and map findings onto generative AI applications.
  • Identify theoretical frameworks that can explain metacognitive processes in human-AI collaboration.
• Experimental Approaches
  • Design creative interventions to support human metacognition before/while using AI and develop experimental protocols to test their effectiveness.
  • Investigate the impacts of using generative AI on human metacognition and cognition(e.g., decision-making, critical thinking, problem-solving).
  • Explore how different AI interface designs (e.g. openAI’s reasoning model) affect users’ metacognitive accuracy.
• Surveys and Interviews
  • Conduct surveys or interviews to investigate factors that determine whether individuals offload cognitive tasks to AI.
  • Explore positive and negative consequences of cognitive offloading to AI.
  • Study domain-specific differences in metacognitive strategies when collaborating with AI.

Starting Date

from August 2025, at least ~6 (Bachelor) / ~8 (Master) months before you plan to hand in and submit your thesis.

Application

Please send a detailed application including your specific topic idea, your CV and transcript of records. For students who are NOT studying WiWi/WINF: Please clarify in advance with the respective degree programme coordinator whether supervision by the chair is permitted. For example, this is often not possible for students from other faculties (TechFak, NatFak).

Requirements

• Interest in interdisciplinary research combining cognitive psychology and AI.
• Basic understanding of experimental design or qualitative research methods.
• Willingness to engage with both technical and psychological literature.
• For experimental approaches: Basic programming skills are beneficial.

References

Clark, A., & Chalmers, D. (1998). The Extended Mind. Analysis, 58(1), 7–19. https://doi.org/10.1093/analys/58.1.7

Dunn, T. L., Gaspar, C., McLean, D., Koehler, D. J., & Risko, E. F. (2021). Distributed metacognition: Increased Bias and Deficits in Metacognitive Sensitivity when Retrieving Information from the Internet. Technology, Mind, and Behavior, 2(3). https://doi.org/10.1037/tmb0000039

Risko, E. F., & Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences, 20(9), 676–688. https://doi.org/10.1016/j.tics.2016.07.002

Tankelevitch, L., Kewenig, V., Simkute, A., Scott, A. E., Sarkar, A., Sellen, A., & Rintel, S. (2024). The Metacognitive Demands and Opportunities of Generative AI. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, 1–24. https://doi.org/10.1145/3613904.3642902

Food systems are a major contributor to global environmental degradation, with dietary choices accounting for roughly 25% of greenhouse gas emissions (Crippa et al., 2021). Despite growing awareness, consumers still struggle to assess the sustainability of food products at the point of sale (Camilleri et al., 2019). This is partly due to the complexity of interpreting sustainability labels and partly due to behavioural barriers. Even when information is available, the intention-behaviour gap often persists (Carrington et al., 2014).

The rise of online grocery shopping allows consumers to make food choices from home (GroceryDoppio, 2023) and opens new opportunities to support more sustainable decisions through targeted interface design (De Bauw et al., 2022). However, real-time sustainability feedback, a potentially powerful intervention, remains underexplored in digital food environments (Shin et al., 2020).

This thesis aims to conceptualize an intervention that combines product-level sustainability information (Clark et al., 2022; Poore & Nemecek, 2018; Ritchie et al., 2022) with real-time, basket-level feedback in an online grocery setting. For example, shoppers could be shown a composite score summarising the overall environmental impact of their basket, based on criteria such as greenhouse gas emissions, land use, or animal welfare. An online experiment should be conducted to test the influence of this feedback on consumer decision-making. If needed, the intervention can be implemented as an interactive prototype or mock-up.

The experiment will be implemented in collaboration with researchers from the Nürnberg Institute for Market Decisions (NIM), who will co-supervise the thesis.

Prerequisites:

• Required:
  • Enrolled Master Student in International Information Systems, International Business Studies, Economics, Marketing, etc. Bachelor students can only be considered upon careful consideration of timeline feasibility.
  • Interest in advancing research in the food domain.
  • Willingness to write their thesis in English.
• Highly Desirable:
  • Previous experience in experimental methodology, e.g., through our courses “Experimentelle Verhaltensforschung in Data Science (EVIDS)” (Bachelor) or the seminar “Information Systems for Behavior Change (ISBC)” (Master).
  • Previous knowledge and/or experience in the application of statistical methods (e.g., ANOVA, LMM, etc.)

References

• Camilleri, A. R., Larrick, R. P., Hossain, S., & Patino-Echeverri, D. (2019). Consumers underestimate the emissions associated with food but are aided by labels. Nature Climate Change, 9(1), 53-58.
• Carrington, M. J., Neville, B. A., & Whitwell, G. J. (2014). Lost in translation: Exploring the ethical consumer intention–behavior gap. Journal of business research, 67(1), 2759-2767.
• Clark, M., Springmann, M., Rayner, M., Scarborough, P., Hill, J., Tilman, D., … & Harrington, R. A. (2022). Estimating the environmental impacts of 57,000 food products. Proceedings of the National Academy of Sciences, 119(33), e2120584119.
• Crippa, M., Solazzo, E., Guizzardi, D., Monforti-Ferrario, F., Tubiello, F. N., & Leip, A. J. N. F. (2021). Food systems are responsible for a third of global anthropogenic GHG emissions. Nature food, 2(3), 198-209.
• De Bauw, M., De La Revilla, L. S., Poppe, V., Matthys, C., & Vranken, L. (2022). Digital nudges to stimulate healthy and pro-environmental food choices in E-groceries. Appetite, 172, 105971.
• GroceryDoppio. (2023, February 7). January 2023: State of Digital Grocery Performance Scorecard. Performance Scorecard. https://www.grocerydoppio.com/performance-scorecard/january-2023-state-of-digital-grocery-performance-scorecard
• Poore, J., & Nemecek, T. (2018). Reducing food’s environmental impacts through producers and consumers. Science, 360(6392), 987-992.
• Ritchie, H., Rosado, P., & Roser, M. (2022). Environmental Impacts of Food Production. Environmental Impacts of Food Production. https://ourworldindata.org/environmental-impacts-of-food
• Shin, S., Van Dam, R. M., & Finkelstein, E. A. (2020). The Effect of Dynamic Food Labels with Real-Time Feedback on Diet Quality: Results from a Randomized Controlled Trial. Nutrients, 12(7), 2158.

Food systems are a major contributor to global environmental degradation, with dietary choices accounting for roughly 25% of greenhouse gas emissions (Crippa et al., 2021). Despite growing awareness, consumers still face difficulties in identifying more sustainable food options during the shopping process (Camilleri et al., 2019). Even when information is available, the complexity of sustainability indicators and behavioural barriers often prevent this knowledge from translating into action (Carrington et al., 2014).

Online grocery platforms offer new opportunities to support sustainable decision-making through digital tools and interface design. Among these tools, recommender systems are widely used to suggest products based on user behaviour or preferences (de Bauw et al., 2022; Jesse & Jannach, 2021). While often optimised for sales, such systems can also be designed to promote sustainable choices (Felfernig et al., 2023).

This thesis aims to conceptualize a sustainability-sensitive recommender system for online grocery environments. The system could prioritise or highlight products with lower environmental impact, based on indicators such as greenhouse gas emissions, land use, or animal welfare (Clark et al., 2022; Poore & Nemecek, 2018; Ritchie et al., 2022). Students are encouraged to explore or propose their own criteria and recommendation logic. An online experiment should be conducted to evaluate the influence of the recommender system on consumer decision-making. If needed, the system can be implemented as an interactive prototype or mock-up.

The experiment will be implemented in collaboration with researchers from the Nürnberg Institute for Market Decisions (NIM), who will co-supervise the thesis.

Prerequisites:

• Required:
  • Enrolled Master Student in International Information Systems, International Business Studies, Economics, Marketing, etc. Bachelor students can only be considered upon careful consideration of timeline feasibility.
  • Interest in advancing research in the food domain.
  • Willingness to write their thesis in English.
• Highly Desirable:
  • Previous experience in experimental methodology, e.g., through our courses “Experimentelle Verhaltensforschung in Data Science (EVIDS)” (Bachelor) or the seminar “Information Systems for Behavior Change (ISBC)” (Master).
  • Previous knowledge and/or experience in the application of statistical methods (e.g., ANOVA, LMM, etc.)

References

• Camilleri, A. R., Larrick, R. P., Hossain, S., & Patino-Echeverri, D. (2019). Consumers underestimate the emissions associated with food but are aided by labels. Nature Climate Change, 9(1), 53-58.
• Carrington, M. J., Neville, B. A., & Whitwell, G. J. (2014). Lost in translation: Exploring the ethical consumer intention–behavior gap. Journal of business research, 67(1), 2759-2767.
• Clark, M., Springmann, M., Rayner, M., Scarborough, P., Hill, J., Tilman, D., … & Harrington, R. A. (2022). Estimating the environmental impacts of 57,000 food products. Proceedings of the National Academy of Sciences, 119(33), e2120584119.
• Crippa, M., Solazzo, E., Guizzardi, D., Monforti-Ferrario, F., Tubiello, F. N., & Leip, A. J. N. F. (2021). Food systems are responsible for a third of global anthropogenic GHG emissions. Nature food, 2(3), 198-209.
• De Bauw, M., De La Revilla, L. S., Poppe, V., Matthys, C., & Vranken, L. (2022). Digital nudges to stimulate healthy and pro-environmental food choices in E-groceries. Appetite, 172, 105971.
• Felfernig, A., Wundara, M., Tran, T. N. T., Polat-Erdeniz, S., Lubos, S., El Mansi, M., … & Le, V. M. (2023). Recommender systems for sustainability: overview and research issues. Frontiers in big Data, 6, 1284511.
• Jesse, M., & Jannach, D. (2021). Digital nudging with recommender systems: Survey and future directions. Computers in Human Behavior Reports, 3, 100052. https://doi.org/10.1016/j.chbr.2020.100052
• Poore, J., & Nemecek, T. (2018). Reducing food’s environmental impacts through producers and consumers. Science, 360(6392), 987-992.
• Ritchie, H., Rosado, P., & Roser, M. (2022). Environmental Impacts of Food Production. Environmental Impacts of Food Production. https://ourworldindata.org/environmental-impacts-of-food

Generative Artificial Intelligence (GenAI) has rapidly become pervasive in our everyday lives. However, seminal studies are already showing a range of negative side-effects of GenAI use on cognitive engagement and abilities, risking cognitive atrophy and overreliance (Kosmyna et al., 2025; Lee et al., 2025; Schoeffer et al., 2025; Zhai et al., 2024). Therefore, it is imperative that human-AI interfaces are designed in a way that promotes deep engagement and the critical reflection of outputs (Yatani et al., 2024).

This thesis is meant to contribute to this endeavor by developing and experimentally testing a design intervention in a (mock-up) LLM interaction interface like ChatGPT that promotes reflection and critical thinking. Students may contribute their own specific ideas for study contexts and settings. Possible interventions could be elements that induce deliberate friction, increase transparency or provide metacognitive scaffolds.

The experiment will be implemented in collaboration with researchers from the Nürnberg Institute for Market Decisions (NIM), who will co-supervise the thesis.

Prerequisites:

• Required:
  • Enrolled Master Student in International Information Systems, International Business Studies, Economics, Marketing, etc. Bachelor students can only be considered upon careful consideration of timeline feasibility.
  • Interest in advancing research in the food domain.
  • Willingness to write their thesis in English.
• Highly Desirable:
  • Previous experience in experimental methodology, e.g., through our courses “Experimentelle Verhaltensforschung in Data Science (EVIDS)” (Bachelor) or the seminar “Information Systems for Behavior Change (ISBC)” (Master).
  • Previous knowledge and/or experience in the application of statistical methods (e.g., ANOVA, LMM, etc.)

References

• Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2506.08872

• Lee, H.-P. (Hank), Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 1–22. https://doi.org/10.1145/3706598.3713778

• Schoeffer, J., Jakubik, J., Vossing, M., Kuhl, N., & Satzger, G. (2025). AI Reliance and Decision Quality: Fundamentals, Interdependence, and the Effects of Interventions. Journal of Artificial Intelligence Research, 82, 471–501.

• Yatani, K., Sramek, Z., & Yang, C.-L. (2024). AI as Extraherics: Fostering Higher-order Thinking Skills in Human-AI Interaction (Version 2). arXiv. https://doi.org/10.48550/ARXIV.2409.09218

• Zhai, C., Wibowo, S., & Li, L. D. (2024). The effects of over-reliance on AI dialogue systems on students’ cognitive abilities: A systematic review. Smart Learning Environments, 11(1), 28. https://doi.org/10.1186/s40561-024-00316-7

Generative Artificial Intelligence (GenAI) has rapidly become pervasive in our everyday lives. However, seminal studies are already showing a range of negative side-effects of GenAI use on user decision-making, risking cognitive atrophy and overreliance (Kosmyna et al., 2025; Lee et al., 2025; Schoeffer et al., 2025; Zhai et al., 2024). However, studies have also shown that carefully designed human-AI interfaces can mitigate such effects and promote deep engagement and the critical reflection of outputs (Yatani et al., 2024).

In this highly relevant emerging research field, an overview of evidence on GenAI overreliance is necessary. The aim of this thesis is to conduct a literature review to investigate which factors can increase or decrease the occurrence of overreliance in human-(Gen)AI interactions. Students may contribute their own ideas to adjusting the scope.

Prerequisites:

Enrolled student at WiSo. If your study program is not a part of WiSo, please check your program requirements whether our team is allowed to supervise your thesis.

References:

• Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2506.08872
• Lee, H.-P. (Hank), Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 1–22. https://doi.org/10.1145/3706598.3713778
• Schoeffer, J., Jakubik, J., Vossing, M., Kuhl, N., & Satzger, G. (2025). AI Reliance and Decision Quality: Fundamentals, Interdependence, and the Effects of Interventions. Journal of Artificial Intelligence Research, 82, 471–501.
• Zhai, C., Wibowo, S., & Li, L. D. (2024). The effects of over-reliance on AI dialogue systems on students’ cognitive abilities: A systematic review. Smart Learning Environments, 11(1), 28. https://doi.org/10.1186/s40561-024-00316-7

Applications are closed for summer semester 2025.BackgroundAccording to the extended mind hypothesis, human cognition extends beyond the brain and nervous system to the body and environmental tools (Clark & Chalmers, 1998). Using technological tools to facilitate cognitive processes is often referred to as “cognitive offloading.” Or “distributed cognition” (Risko & Gilbert, 2016). In the age of generative AI and its increasing capabilities, the possibilities for humans to offload energy- and time-consuming cognitive processes to AI are numerous and evolving rapidly.Research Gap

In many human-AI collaboration scenarios, humans remain the “final authority” to accept or reject AI recommendations. Therefore, using AI for task completion not only necessitates controlling and monitoring one’s own cognitive processes (often referred to as metacognition) but also evaluating AI’s processes and outputs (Dunn et al., 2021; Tankelevitch et al., 2024). Despite the increasing metacognitive demands of generative AI and the relevance of accurate metacognition to prevent under-/overreliance on AI outputs, there is a lack of research investigating the role of human metacognition in (successful) Human-AI collaborations.

Thesis Goals

Bachelor’s and Master’s theses on this topic aim to investigate the dual role of metacognition: first, in the decision-making process of when to offload cognitive tasks to AI; and second, in how humans evaluate AI outputs once collaboration occurs. Research should explore how targeted interventions can be designed to enhance metacognitive accuracy during human-AI collaboration. The goal is to gain conceptual and empirical insights that will advance our understanding of human-AI cognitive partnerships and inform future design approaches.

Research Approaches

• (Systematic) Literature Review
  • Conduct a (systematic) literature search exploring interventions (such as cognitive forcing strategies) that support metacognitive monitoring/control and prevent excessive cognitive offloading.
  • Review interdisciplinary research and map findings onto generative AI applications.
  • Identify theoretical frameworks that can explain metacognitive processes in human-AI collaboration.
• Experimental Approaches
  • Design creative interventions to support human metacognition before/while using AI and develop experimental protocols to test their effectiveness.
  • Investigate the impacts of using generative AI on human metacognition and cognition(e.g., decision-making, critical thinking, problem-solving).
  • Explore how different AI interface designs (e.g. openAI’s reasoning model) affect users’ metacognitive accuracy.
• Surveys and Interviews
  • Conduct surveys or interviews to investigate factors that determine whether individuals offload cognitive tasks to AI.
  • Explore positive and negative consequences of cognitive offloading to AI.
  • Study domain-specific differences in metacognitive strategies when collaborating with AI.

Starting Date

from August 2025, at least ~6 (Bachelor) / ~8 (Master) months before you plan to hand in and submit your thesis.

Application

Please send a detailed application including your specific topic idea, your CV and transcript of records. For students who are NOT studying WiWi/WINF: Please clarify in advance with the respective degree programme coordinator whether supervision by the chair is permitted. For example, this is often not possible for students from other faculties (TechFak, NatFak).

Requirements

• Interest in interdisciplinary research combining cognitive psychology and AI.
• Basic understanding of experimental design or qualitative research methods.
• Willingness to engage with both technical and psychological literature.
• For experimental approaches: Basic programming skills are beneficial.

References

Clark, A., & Chalmers, D. (1998). The Extended Mind. Analysis, 58(1), 7–19. https://doi.org/10.1093/analys/58.1.7

Dunn, T. L., Gaspar, C., McLean, D., Koehler, D. J., & Risko, E. F. (2021). Distributed metacognition: Increased Bias and Deficits in Metacognitive Sensitivity when Retrieving Information from the Internet. Technology, Mind, and Behavior, 2(3). https://doi.org/10.1037/tmb0000039

Risko, E. F., & Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences, 20(9), 676–688. https://doi.org/10.1016/j.tics.2016.07.002

Tankelevitch, L., Kewenig, V., Simkute, A., Scott, A. E., Sarkar, A., Sellen, A., & Rintel, S. (2024). The Metacognitive Demands and Opportunities of Generative AI. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, 1–24. https://doi.org/10.1145/3613904.3642902

Background

Advances in computational power and declining data-processing costs have accelerated the diffusion of artificial intelligence (AI) across organizations, as firms increasingly implement AI-based systems with the expectation of enhancing performance and welfare (Glikson & Wilson, 2023; Ludwig & Achtziger, 2023). Despite these expectations, prior studies suggest that the anticipated performance gains from AI adoption often fail to materialize, leaving substantial welfare potential untapped (Vaccaro et al., 2024; De Freitas et al., 2023). This has shifted scholarly attention toward the human side of AI deployment, raising the question of the conditions under which decision makers disregard or resist AI-based decision support.

A central concept in this literature is algorithm aversion, which refers to individuals’ tendency to prefer human judgment over algorithmic advice (Burton et al., 2020; Mahmud et al., 2022; Jussupow et al., 2020). Dietvorst et al. (2015) initially attributed algorithm aversion to individuals’ heightened sensitivity to algorithmic errors compared to human errors. Subsequent research has identified additional drivers, including concerns that algorithms fail to account for individual circumstances (Longoni et al., 2019) and resistance in domains perceived as subjective or intuition-based (Castelo et al., 2019). At the same time, other studies highlight algorithm appreciation, where individuals value and rely on algorithmic input, resulting in mixed and sometimes contradictory empirical findings (Logg et al., 2019). This heterogeneity complicates the derivation of robust conclusions about when and why algorithm aversion occurs.

One plausible explanation for these inconsistencies lies in methodological variation across studies. Prior research differs in the operationalization of dependent variables (Zehnle et al., 2024), the use of hypothetical versus real decision contexts (Logg & Schlund, 2024), and the conceptualization of the human–AI relationship (Jussupow et al., 2024). Consequently, algorithm aversion may be captured in ways that are not fully comparable across studies, potentially producing apparent contradictions driven more by measurement and design choices than by substantive differences.

Thesis Goals

Against this background, this thesis aims to provide a descriptive methodological review of the algorithm aversion literature. Following a structured search and screening process, the review will systematically code and analyze the methodological characteristics of existing studies. The objectives are to map how algorithm aversion is operationalized and measured, identify dominant study designs, and uncover methodological blind spots or underexplored areas. By clarifying how algorithm aversion has been studied to date, the review seeks to facilitate the synthesis of existing findings and inform methodological decisions in future research.

Methodological approach

• Descriptive Literature Review: What methodological limitations and underexplored areas can be identified in the algorithm aversion literature?

Starting Date

As of now.

Requirements

• Interest in new technologies and user-centric perspectives, particularly human–AI interaction.
• Good (!) knowledge of research methods and the ability to distinguish between common approaches (e.g., surveys, experiments, interviews, and field studies).

Support provided

• Students will receive access to core literature on algorithm aversion.
• Students will receive access to “how to” literature regarding descriptive reviews.
• Students will receive a predefined search string for their review as well as guidance on screening and coding procedures.

Application

Please send a detailed application including your specific topic idea, your CV and transcript of records. For students who are NOT studying WiWi/WINF: Please clarify in advance with the respective degree programme coordinator whether supervision by the chair is permitted. For example, this is often not possible for students from other faculties (TechFak, NatFak).

References