For additional information about the M&O Seminars, please contact Shelly Whitmer at [email protected].
monday, JANUARY 20
Jerry Davis, Ross M&O
Title: AI in Management Research: Revolutionary Tool or Existential Risk
Abstract: Jerry will deliver a talk and introduce us to the most concerning aspects of GAI, highlighting the implications and potential dilemmas we may face.
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, January 27
Arik Cheshin, UM Visiting Professor
Title: Regulating and Emerging: How Regulating Emotions of Others Iimpacts Selection to a Special Elite Military Unit
Abstract: The ability to influence and improve others' emotions is increasingly recognized as a central component of effective leadership. But does enhancing the emotional state of others contribute to being identified as an informal leader or being deemed fit for inclusion in elite military units? Established research acknowledges the role of emotions in leadership, and recent findings suggest that improving others' emotions can lead to being selected as an informal leader (Cheshin & Luria, 2024). This study extends that research into a field setting involving 106 recruits undergoing tryouts for a specialized elite army unit. Surveys were conducted at three intervals: the start of the tryouts session, 10 hours in, and 24 hours later. The recruits were divided into six teams, and by the end of the session, 83 candidates remained, with 20 being selected by military recruiters. Our methodology included collecting self-reported data on emotion regulation and peer evaluations of who influenced team emotions and who was considered the informal leader (excluding self-nominations). We then compared these peer evaluations to the formal assessments made by military recruiters. This presentation will discuss whether the regulation of others' emotions—specifically extrinsic affect improvement—is a coveted skill for informal leadership and being selected to an elite military unit. I will share our initial findings and look forward to your insights and feedback.
Time: 11:45 a.m. - 12:45 p.m.
Location: Blau Colloquium
monday, February 3
Andrei Boutyline, UM Sociology
Title: Meaning in Hyperspace: Word Embeddings as Models of Culture
Abstract: Word embeddings are language models that represent words or concepts as positions in an abstract many-dimensional meaning space. Despite a growing range of applications demonstrating their utility for social science, there is little conceptual clarity regarding what exactly embeddings measure and whether this matches what we need them to measure. I fill this theoretical gap by arguing that embeddings operationalize context spaces, where words’ positions can reflect any regularity in usage. Most sociological scholarship, however, is interested in concept spaces, where positions strictly indicate meaningful conceptual features (e.g., femininity or status). Because meaningful features yield regularities in usage, context spaces can proxy for concept spaces. However, context spaces also reflect regularities in the surface form of language—e.g., syntax, morphology, and dialect—which are irrelevant to most sociological investigations and can bias cultural measurement. I draw on our framework to propose best practices for successfully measuring meaning with embeddings.
Time: 11:45 a.m. - 12:45 p.m.
Location: Blau Colloquium
monday, february 10
Amie Gordon, UM Psychology
Title: Can we Crack the Code? Leveraging Modern Technology to Better Predict Relationship Formation and Dissolution
Abstract: Despite decades of research, relationship scientists have yet to fully unpack the mysteries of dating and compatibility. We still cannot robustly predict the formation and dissolution of any given romantic relationship. Furthermore, the researchers who study attraction and initial dating tend to have little overlap with those who study the maintenance of established relationships, leaving many unanswered questions about trajectories of relationships over time. In this talk, I will discuss two ongoing projects that attempt to fill some of the gaps in this literature by leveraging modern technology to gather large-scale data on the formation and dissolution of romantic relationships. One project uses intensive longitudinal methods to track moment-to-moment relationship experiences with the goal of better understanding the dynamic patterns of change in relationship quality and stability over time. The second project leverages dating apps—the most common way for couples to meet today—to gain new insights into relationship formation. In collaboration with Elizabeth Bruch, a U-M sociologist, my lab has spent the past two years developing Revel, a dating app designed for science, not profit. This talk showcases how these projects might enable us to track the full lifecycle of relationships at an unprecedented scale and hopefully shed light on whether relationship scientists will ever be able to confidently invest their retirement savings in predicting which people will “click” and which relationships will endure.
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, february 17
Douglas Guilbeault, Stanford
Title: Large-Scale Evidence of Gendered Ageism in Online Data and Workplace AI
Abstract: Policy reports, media coverage, and workplace interviews reveal that older women face a dual bias against both their gender and age, negatively impacting not only their mental health, but also their professional status and opportunities. Yet strikingly little is known about the online prevalence of this dual bias and its broader impact. Here, we show that age-based gender bias pervades the internet through images, videos, and textual data, with detrimental effects on both people and artificial intelligence. First, we examine the age representation of women and men across 3,435 social categories (e.g., “doctor” or “friend”) in over 1.3 million recent images from Google, Wikipedia, IMDb, and Flickr, alongside thousands of YouTube videos. Across datasets and measurement strategies, women are consistently represented as younger than men. Next, we conducted a nationally representative, pre-registered experiment showing that searching for Google images of occupations amplifies bias in people’s beliefs about the age of women and men, as well as their preference for hiring younger women and older men. We then generalize beyond visual content by showing that women and men are represented as younger and older, respectively, in nine language models trained on billions of words from the internet. We conclude by showing how one of these language models, ChatGPT, perpetuates age-based gender bias in workplace applications. When generating resumes, ChatGPT assumes women are younger and less experienced than men, and it rates resumes from older applicants — and especially men — as higher quality. These findings capture the multimodal and multidimensional nature of gender bias, revealing unique challenges and opportunities in the fight against gender inequality online.
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, march 10
Robin Edelstein, UM Psychology
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, march 17
Catherine Shea, CMU Tepper
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, march 24
Rick Larrick, Duke
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, march 31
TBD
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, april 7
Ray Reagans, MIT
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, april 14
Jordan Nye, Ross M&O
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons
monday, april 21
Anusha Kallapur, Ross M&O
Title: TBD
Abstract: TBD
Time: 11:45 a.m. - 12:45 p.m.
Location: B1590 Corner Commons