Verfügbare Themen

Supervisor: Claire Schneiders 

Level: Bachelor  

Background 

Artificial Intelligence (AI) is increasingly used to support organizational activities such as customer service, sales, software development, and decision-making. These applications can create business value through higher productivity, lower costs, increased revenues, improved quality, or innovation. Existing research provides frameworks for understanding these value-generating mechanisms (Enholm et al., 2022). 

Empirical studies also document measurable outcomes in organizational practice. For example, AI assistance has increased issues resolved per hour in customer support (Brynjolfsson et al., 2025), while firm-level research links AI investments to sales growth and product innovation (Babina et al., 2024). However, improvements in task performance do not automatically translate into lower organizational costs or higher profits. Assessing realized business value therefore requires attention to what is measured, how it is measured, and whether the costs of implementing and operating AI are considered. 

Research Gap 

Existing reviews address AI adoption, value-generating mechanisms, and firm-level economic effects (Enholm et al., 2022; Babina, 2026). Building on this literature, this thesis focuses specifically on quantitative evidence of realized outcomes and the strength of the claims that can be drawn from it. 

An important unresolved issue is how evidence from individual tasks and specific applications relates to business value at the process and firm levels. In particular, a structured assessment is needed of: 

• which findings concern performance in experimental tasks and which document outcomes in actual organizational operations; 

• whether reported benefits are operational improvements, observed financial outcomes, or estimated monetary equivalents; 

• how convincingly the observed outcomes can be attributed to AI rather than other organizational changes; and 

• whether reported gains remain economically meaningful after implementation, operating, and complementary organizational costs are considered. 

Thesis Goals 

This thesis aims to systematically identify, compare, and critically assess empirical evidence on the realized business value of AI. It will develop an evidence map showing where measurable value has been documented, through which mechanisms, and with what limitations. 

You may focus on the following research questions: 

• What quantitative evidence exists on the realized business value of AI applications in organizations? 

• In which organizational functions and use cases have economically meaningful outcomes been demonstrated, and where are effects limited, absent, or negative? 

• Through which mechanisms does AI contribute to productivity, costs, revenues, quality, or innovation? 

• At which level is value measured (task, employee, process, or firm) and how robust is the underlying evidence? 

• Which characteristics of tasks, employees, AI applications, and organizational contexts are associated with differences in outcomes? 

Methodological Approach 

The thesis will follow a systematic literature review approach. Students will develop a review protocol, define transparent inclusion and exclusion criteria, search relevant academic databases, and document study selection in line with PRISMA 2020 reporting guidance (Page et al., 2021). The findings will be synthesized in an evidence map and a structured narrative analysis.  

Requirements 

• Willingness to write the thesis in English. 

• Interest in the intersection of AI, information systems, and organizational performance. 

• Ability and willingness to independently search, organize, and critically evaluate academic literature. 

• Basic knowledge of quantitative research methods and statistical interpretation; experience with systematic reviews is an advantage. 

Selected Literature 

Enholm, I. M., Papagiannidis, E., Mikalef, P., & Krogstie, J. (2022). Artificial Intelligence and Business Value: a Literature Review. Information Systems Frontiers, 24, 1709–1734. https://doi.org/10.1007/s10796-021-10186-w 

Babina, T., Fedyk, A., He, A., & Hodson, J. (2024). Artificial intelligence, firm growth, and product innovation. Journal of Financial Economics, 151, 103745. https://doi.org/10.1016/j.jfineco.2023.103745 

Brynjolfsson, E., Li, D., & Raymond, L. (2025). Generative AI at Work. The Quarterly Journal of Economics, 140(2), 889–942. https://doi.org/10.1093/qje/qjae044 

Babina, T. (2026). Understanding Firms’ AI Efforts and Their Economic Impact. CEPR Discussion Paper 21804; also NBER Working Paper 35123. Working paper. https://cepr.org/publications/dp21804 

Page, M. J., et al. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71 

Supervisor: Claire Schneiders 

Level: Bachelor or Master 

Background 

Providing feedback and coaching is a central responsibility of sales managers. Managers help salespeople interpret their performance, prioritize activities, identify improvement opportunities, and work toward individual sales goals. 

At the same time, organizations increasingly use algorithmic and AI-based systems to perform similar functions. Based on customer relationship management (CRM) data, such systems can process large amounts of behavioral and performance data and provide personalized recommendations at scale. 

Research shows that AI-generated performance feedback can increase the accuracy and consistency of information analysis and provide feedback that is more relevant to an individual employee’s situation (Tong et al., 2021). However, employees do not necessarily respond to algorithmic and human feedback in the same way. AI-generated feedback has been associated with lower feedback acceptance, lower motivation, lower acceptance of the feedback provider, and lower intentions to seek further feedback compared with human feedback (Hein et al., 2024). 

Research Gap 

Previous research has mainly compared how employees respond to feedback or coaching provided by an AI system versus a human manager (Tong et al., 2021; Hein et al., 2024; Casenave & Schmitt, 2025). 

However, there is still limited knowledge about what happens when employees receive feedback from both sources and can compare the usefulness of algorithmic feedback with the usefulness of their manager’s feedback. 

In particular, it is unclear: 

• whether useful algorithmic feedback changes employees’ perceptions of their manager’s competence, 

• whether the usefulness of the manager’s own feedback influences this relationship, 

• and whether the combination of both feedback sources affects employees’ willingness to seek advice from their manager in the future. 

Thesis Goals 

This thesis aims to investigate how the usefulness of algorithmic and managerial feedback jointly shapes employees’ perceptions of their sales manager. 

You may focus on one of these research questions: 

• How does the usefulness of algorithmic feedback affect employees’ perceived competence of their sales manager? 

• Does the usefulness of managerial feedback moderate the effect of algorithmic feedback on perceived managerial competence? 

• How do algorithmic and managerial feedback jointly affect employees’ willingness to seek further advice from their manager? 

• Does perceived managerial competence explain how different combinations of algorithmic and managerial feedback influence employees’ willingness to seek managerial advice? 

Methodological Approach 

The thesis will follow a quantitative experimental research approach based on a 2×2 between-subjects online experiment. 

Students will review the relevant literature, develop and pretest the experimental scenario, implement the online experiment and random assignment, conduct an a priori power analysis, independently recruit suitable participants, and analyze main and interaction effects using RStudio. 

Requirements 

• Willingness to write the thesis in English. 

• Ability and willingness to independently recruit suitable participants, ideally sales professionals, for example through personal or professional networks or LinkedIn. 

• Interest in the intersection of leadership, artificial intelligence, and algorithmic management. 

• Basic knowledge of experimental research methods and quantitative data analysis. 

• Willingness to work with R / RStudio. 

Selected Literature 

Casenave, E., & Schmitt, L. (2025). Comparing AI coaching and sales manager coaching: A construal-level approach. Journal of Business Research, 190, 115241. https://doi.org/10.1016/j.jbusres.2025.115241 

Hein, I., Cecil, J., & Lermer, E. (2024). Acceptance and motivational effect of AI-driven feedback in the workplace: An experimental study with direct replication. Frontiers in Organizational Psychology, 2, 1468907. https://doi.org/10.3389/forgp.2024.1468907 

Tong, S., Jia, N., Luo, X., & Fang, Z. (2021). The Janus face of artificial intelligence feedback: Deployment versus disclosure effects on employee performance. Strategic Management Journal, 42(9), 1600–1631. https://doi.org/10.1002/smj.3322 

Supervisor: Claire Schneiders 

Level: Bachelor  

Background 

Sales organizations increasingly use customer relationship management (CRM) systems to document sales activities and track progress toward individual sales goals. Based on these data, algorithmic systems can provide personalized feedback and recommendations that help employees assess their performance and identify ways to improve it (Tong et al., 2021). 

However, such feedback can also raise concerns about monitoring and control. Employees may perceive data-driven feedback as intrusive or stressful, particularly when it feels controlling rather than developmental (Wells et al., 2007; Ravid et al., 2023). 

This creates an important question for organizations using data-driven feedback systems: When does such feedback feel like helpful coaching, and when is it experienced as a form of surveillance? 

Research Gap 

Existing research has investigated employees’ reactions to algorithmic feedback and electronic performance monitoring. Studies suggest that algorithmic feedback can provide useful, individualized performance information, while employees’ reactions depend partly on how they perceive the technology and its purpose (Tong et al., 2021; Hein et al., 2024). Research on electronic performance monitoring further suggests that employees respond differently depending on how monitoring is implemented and perceived (Wells et al., 2007; Ravid et al., 2023). 

However, less is known about how employees experience recurring, personalized feedback based on their own CRM activity data and linked to a self-set sales goal. 

In particular, it remains unclear: 

• which characteristics of CRM-based feedback make it feel helpful rather than controlling, 

• why different employees interpret the same type of feedback differently, 

• and how perceived support versus surveillance influences employees’ willingness to engage with the feedback. 

Thesis Goals 

This thesis aims to explore when and for whom CRM-based algorithmic feedback is perceived as supportive coaching rather than employee surveillance. 

You may focus on one of these research questions: 

• How do employees experience recurring, personalized algorithmic feedback? 

• When is such feedback perceived as helpful and supportive, and when as controlling or intrusive? 

• Which factors may explain differences in employees’ reactions? 

• Under which conditions are employees willing to engage with and act on algorithmic feedback? 

Methodological Approach 

The thesis will combine a structured literature review with a qualitative interview study. 

First, students will review relevant research on algorithmic feedback and management, electronic performance monitoring, employee surveillance, and developmental versus controlling forms of feedback. The review should identify relevant concepts, theories, and potential explanations for differences in employees’ perceptions. 

Second, students will conduct semi-structured interviews with employees who have experience with CRM-based or comparable data-driven feedback systems. The interviews should explore how employees interpret such feedback, what they perceive as helpful or intrusive, and which factors shape their reactions. 

Students will develop an interview guide based on the literature review, independently recruit suitable interview participants, conduct and transcribe the interviews, and analyze the collected data using an established qualitative analysis method. 

Requirements 

• Willingness to write the thesis in English. 

• Ability and willingness to independently recruit suitable interview participants, ideally employees with experience using CRM systems or receiving data-driven performance feedback. 

• Interest in the intersection of information systems, artificial intelligence, and employee monitoring. 

• Interest in qualitative research and conducting interviews. 

• Basic knowledge of qualitative research methods is an advantage. 

Selected Literature 

Hein, I., Cecil, J., & Lermer, E. (2024). Acceptance and motivational effect of AI-driven feedback in the workplace: An experimental study with direct replication. Frontiers in Organizational Psychology, 2, 1468907. 

Ravid, D. M., White, J. C., Tomczak, D. L., Miles, A. F., & Behrend, T. S. (2023). A meta-analysis of the effects of electronic performance monitoring on work outcomes. Personnel Psychology, 76(1), 5–40. 

Tong, S., Jia, N., Luo, X., & Fang, Z. (2021). The Janus face of artificial intelligence feedback: Deployment versus disclosure effects on employee performance. Strategic Management Journal, 42(9), 1600–1631. 

Wells, D. L., Moorman, R. H., & Werner, J. M. (2007). The impact of the perceived purpose of electronic performance monitoring on an array of attitudinal variables. Human Resource Development Quarterly, 18(1), 121–138.

Supervisor: Tim Zehner

Students often avoid seeking help on academic tasks even when they need it. The academic help-seeking literature shows that help-seeking is a public behavior that can signal lower competence relative to peers, and that a performance-avoidance goal orientation increases this avoidance, while a mastery orientation supports adaptive help-seeking (Ryan, Pintrich & Midgley, 2001). Research on visibility and anonymity confirms that students deliberately choose private over public channels when they feel less competent, and that anonymous tools measurably reduce the social anxiety associated with asking questions in large courses.

In parallel, research on AI tutors shows that students partly prefer AI sources for convenience and lower perceived social risk, and a recent study correlates course size (as a proxy for anonymity) with reduced social risk in AI-assisted help-seeking – but without experimentally manipulating visibility. No study to date combines source (AI vs. human) and visibility in a controlled experiment with actual behavior as the dependent variable. This thesis closes exactly that gap.

Potential research questions include:

• Do students seek feedback more frequently when the source is an AI tutor rather than a human instructor/TA?
• Does this effect depend on whether the interaction is visible to classmates versus private?
• Does perceived help-seeking threat mediate the relationship between feedback source and feedback-seeking behavior?
• Does the effect differ depending on students‘ achievement goal orientation (mastery vs. performance-avoidance)?

Methodologically, the thesis would conduct an online experiment with university students (e.g., via Prolific) with a 2×2 between-subjects design (feedback source: AI tutor vs. human instructor/TA × visibility: private vs. visible to peers), holding the interface constant across conditions and measuring actual feedback-seeking behavior as the primary dependent variable, with help-seeking threat as a mediator and goal orientation as a moderator.

Within a bachelor’s thesis, only the experimental concept is developed and elaborated, without actually collecting or analyzing data. A master’s thesis based on the same concept would additionally run the experiment and analyze the data.

Supervisor: Tim Zehner

Feedback-seeking behavior (the deliberate effort employees make to obtain evaluative information about their performance) has long been studied as a driver of learning, self-regulation and performance improvement. Ashford and Cummings‘ foundational work shows that this effort is shaped not only by the expected value of feedback, but by its perceived costs, particularly ego costs and image costs. The growing presence of AI-based feedback sources (chatbots, AI coaches, automated performance dashboards) plausibly changes this cost-benefit calculation, yet it remains unclear how.

This thesis systematically reviews empirical evidence on how the availability of AI as a feedback source affects employees‘ feedback-seeking behavior, compared to seeking feedback from human sources.

The thesis aims to map where empirical evidence on AI-related feedback-seeking is strong, through which theoretical mechanisms effects are explained, and where the literature still relies on self-report or hypothetical scenarios rather than observed behavior.

Potential research questions include:

• What empirical evidence exists on differences in feedback-seeking frequency, method, and content when AI versus human sources are available?
• Through which mechanisms (image costs, ego costs, perceived expertise, trust) does AI availability change feedback-seeking behavior?
• Under which conditions (visibility to others, task type, feedback valence) are these effects strongest or weakest?
• Which theoretical frameworks (feedback-seeking theory, self-determination theory, trust-in-automation research) have been used, and where do they converge or conflict?

Methodologically, the thesis would conduct a systematic literature review across organizational behavior, HCI, and information systems venues, and develop a synthesis of findings and open gaps.

Supervisor: Claire Schneiders 

Level: Bachelor or Master 

Background 

In B2B sales, salespeople typically work on multiple potential deals at the same time.  These opportunities progress through different stages, ranging from initial customer contact and qualification to proposals, negotiations, and contract closure. Together, these opportunities form the sales pipeline (Söhnchen & Albers, 2010). 

Since only some of these opportunities will ultimately result in a sale and salespeople have limited time and resources, they need to decide where to focus their effort. One way to support such prioritization is to estimate the probability of winning an opportunity. However, accurately assessing such probabilities can be challenging, as human forecasts may be affected by biases such as optimism (Kahneman & Lovallo, 1993). 

Information systems can provide different forms of support for estimating win probabilities, including: 

• CRM systems that document the progress of an opportunity, such as customer interactions, meetings, or proposals 
• predefined win probabilities associated with different sales stages 
• data-driven or machine learning-based predictions (Syam & Sharma, 2018) 

At the same time, salespeople may form their assessments largely based on their experience, intuition, or informal judgment. 

Research Gap 

Although sales pipeline management and predictive sales analytics have received considerable attention, there is still limited knowledge about how win probabilities are assessed and used in day-to-day B2B sales practice. 

In particular, it is unclear: 

• how salespeople and managers actually arrive at their own estimates, 
• whether and how these estimates are documented, 
• what forms of system support are available and actually used, and 
• how system-based information and personal judgment interact. 

Thesis Goals 

This thesis aims to investigate how B2B salespeople assess the likelihood of winning sales opportunities and to what extent information systems support this decision-making process. 

You may focus on one of these research questions: 

• How do salespeople and managers form their own win probability estimates, and do they explicitly track them? 
• What types of system support are available in their organizations, such as CRM-based opportunity tracking, stage-specific probabilities, or ML-based predictions? 
• How do salespeople and managers use this information in their decision-making? 
• To what extent are win probabilities used to prioritize opportunities and allocate their time and effort? 

Methodological Approach 

The thesis will follow a qualitative research approach based on semi-structured interviews with salespeople and/or sales managers from different B2B companies (Myers & Newman, 2007). Students will develop an interview guide, recruit suitable interview partners, conduct and transcribe the interviews, and analyze the collected data using an established qualitative analysis method. 

Requirements 

• Willingness to write the thesis in English. 
• Ability and willingness to independently recruit suitable interview partners, e.g. through personal or professional networks or LinkedIn. 
• Interest in the intersection of information systems, decision-making, CRM, and B2B sales. 
• Previous knowledge of sales is not required; students should be willing to familiarize themselves with the basics of B2B sales and sales pipeline management. 
• Basic knowledge of qualitative research methods is an advantage. 

Please note: Applications are closed.

Supervisor: Leonie Manzke

Level: Master

Timeframe: Start possible from April 2026; Submission possible by October 2026 at the earliest

When consumers use AI search summaries to inform purchase decisions, they are often presented with compact, polished guidance that highlights certain claims while deemphasizing others. As a result, users may rely on the summary’s most salient information rather than independently verifying the underlying sources. This makes AI summary contexts a strong candidate for studying trust calibration: whether people appropriately accept high-quality information while rejecting dubious or misleading claims (source).

Research Gap

While trust in AI and credibility assessment have been studied across many contexts, there is limited experimental evidence on how specific interface elements in AI search summaries shape (a) the assessment of credibility of claims within such a summary, (b) detection of dubious content, and (c) alignment between the information provided and subsequent opinion formation and behavior. In particular, interface cues that suggest breadth or consensus (for example, a “source count” label) or early attention-demanding cues (for example, an entry-sentence verdict) may shift which claims dominate advice decisions, potentially increasing overreliance when dubious claims are present.

Thesis Goals

This Master’s thesis aims to test how interface elements in a realistic, online AI Overview-style summary influence trust calibration and decision-making in a consumer advice task (= advising a fictional neighbor what to do).

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

Requirements

• Strong understanding of experimental design, measurement, and statistics.
• It is highly recommended you have taken either of our courses “Experimentelle Verhaltensforschung in Data Science” or “IS for Behavior Change”.
• Basic programming skills for stimulus implementation and click logging are not essential, but beneficial.

Please note: Applications are closed.

Level: Bachelor

Timeframe: Start possible from April 2026; Submission possible by August 2026 at the earliest

Supervisor: Leonie Manzke

When consumers seek information online to inform a purchase decision, they often aim to quickly extract actionable guidance rather than evaluate every available source in depth. AI-assisted search and AI-generated summaries have reshaped this process by presenting compact, synthesized claims in polished language, which many users treat as a “one-stop solution” (Kaiser et al., 2025). As a result, decision-making now relies on a fundamentally changed process for judging the credibility of information. Users must decide which summary claims to accept or discount, and translate a mix of credible and dubious information into judgments.

Research Gap

Credibility research is well established, but past work has focused on the credibility of websites, social media content, customer reviews, or traditional online search (Fogg, 2003; Suarez-Lledo & Alvarez-Galvez, 2021; Cheung, Sia & Kuan, 2012; Brand & Reith, 2022; Kammerer & Gerjets, 2012). Since AI summaries have become ubiquitous, the process of consumer information-seeking has been fundamentally transformed. Therefore, there is a need for an integrative framework that organizes which antecedents are likely to shape credibility assessments of claims found in AI summaries, and how these antecedents may operate during online information-seeking.

Thesis Goals

This Bachelor’s thesis aims to conduct a scoping literature review (Arksey & O’Malley, 2005; Levac et al., 2010) in order to develop a conceptual framework that explains credibility assessments of claims presented in AI-generated summaries in online information-seeking processes.

Requirements

• Willingness to write the thesis in English.
• Interest in interdisciplinary work (consumer judgment, information behavior, credibility research, human–AI interaction).
• Basic familiarity with empirical research is beneficial.

References

Arksey, H., & O’Malley, L. (2005). Scoping studies: Towards a methodological framework. International Journal of Social Research Methodology, 8(1), 19–32. https://doi.org/10.1080/1364557032000119616

Brand, B. M., & Reith, R. (2022). Cultural differences in the perception of credible online reviews – The influence of presentation format. Decision Support Systems, 154, 113710. https://doi.org/10.1016/j.dss.2021.113710

Cheung, C. M. K., Sia, C. L., & Kuan, K. K. Y. (2012). Is this review believable? A study of factors affecting the credibility of online consumer reviews from an ELM perspective. Journal of the Association for Information Systems, 13(8), 618–635. https://doi.org/10.17705/1jais.00305

Fogg, B. J. (2003). Prominence-interpretation theory: Explaining how people assess credibility online. In CHI ’03 extended abstracts on Human factors in computing systems (p. 722). https://doi.org/10.1145/765891.765951

Kaiser, C., Kaiser, J., Schallner, R., & Schneider, S. (2025). A new era of online search? A large-scale study of user behavior and personal preferences during practical search tasks with generative AI versus traditional search engines. In Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (pp. 1–7). https://doi.org/10.1145/3706599.3720123

Kammerer, Y., & Gerjets, P. (2012). How search engine users evaluate and select web search results: The impact of the search engine interface on credibility assessments. In D. Lewandowski (Ed.), Web engine research (pp. 205–233). Emerald Group Publishing Limited. https://doi.org/10.1108/S1876-0562(2012)002012a012

Levac, D., Colquhoun, H., & O’Brien, K. K. (2010). Scoping studies: Advancing the methodology. Implementation Science, 5(1), Article 69. https://doi.org/10.1186/1748-5908-5-69

Suarez-Lledo, V., & Alvarez-Galvez, J. (2021). Prevalence of health misinformation on social media: Systematic review. Journal of Medical Internet Research, 23(1), Article e17187. https://doi.org/10.2196/17187

Applications are closed.

Supervisor: Sophie Kuhlemann

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.

Level: Master or Bachelor, provided requirements are fulfilled.

Methodological approach:

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

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.

References

Burton, J. W., Stein, M., & Jensen, T. B. (2020). A systematic review of algorithm aversion in augmented decision making. Journal of Behavioral Decision Making, 33(2), 220–239. https://doi.org/10.1002/bdm.2155

Castelo, N., Bos, M. W., & Lehmann, D. R. (2019). Task-Dependent Algorithm Aversion. Journal of Marketing Research, 56(5), 809–825. https://doi.org/10.1177/0022243719851788

Dietvorst, B. J., Simmons, J. P., & Massey, C. (2015). Algorithm aversion: People erroneously avoid algorithms after seeing them err.

Glikson, E., & Woolley, A. W. (2020). Human Trust in Artificial Intelligence: Review of Empirical Research.

Jussupow, E., Benbasat, I., & Heinzl, A. (2024). An Integrative Perspective on Algorithm Aversion and Appreciation in Decision-Making.

Jussupow, E., Benbasat, I. J., & Heinzl, A. (2020, Juni 15). Why are we averse towards algorithms? A comprehensive Literature Review on Algorithm Aversion.

Logg, J. M., Minson, J. A., & Moore, D. A. (2019). Algorithm appreciation: People prefer algorithmic to human judgment. Organizational Behavior and Human Decision Processes, 151, 90–103. https://doi.org/10.1016/j.obhdp.2018.12.005

Logg, J., & Schlund, R. (2024). A simple explanation reconciles “algorithm aversion” and “algorithm appreciation”: Hypotheticals vs. Real judgments.

Longoni, C., Bonezzi, A., & Morewedge, C. K. (2019). Resistance to Medical Artificial Intelligence. Journal of Consumer Research, 46(4), 629–650. https://doi.org/10.1093/jcr/ucz013

Ludwig, J., & Achtziger, A. (2021). Cognitive misers on the web: An online-experiment of incentives, cheating, and cognitive reflection. Journal of Behavioral and Experimental Economics, 94. Scopus. https://doi.org/10.1016/j.socec.2021.101731

Mahmud, H., Islam, A. K. M. N., Ahmed, S. I., & Smolander, K. (2022). What influences algorithmic decision-making? A systematic literature review on algorithm aversion. Technological Forecasting and Social Change, 175, 121390. https://doi.org/10.1016/j.techfore.2021.121390

Zehnle, M., Hildebrand, C., & Valenzuela, A. (2025). Not all AI is created equal: A meta-analysis revealing drivers of AI resistance across markets, methods, and time. International Journal of Research in Marketing, S0167811625000114. https://doi.org/10.1016/j.ijresmar.2025.02.005

Please note: Applications are closed.

Supervisor: Laura Schneider

According 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.

Level: Bachelor or Master

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.

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