Salini
Pillai
Generative Gender Gaps: Evidence from LLM Summarization
Abstract profile. Full document pending author claim.
Authors:
Salini Pillai, Rembrand Koning, Tarun Khanna
Date Created:
2025-01-01
Course Title:
Professor:
Not specified
About Paper:
In the modern day, generative AI tools have rapidly reshaped evaluate the likelihood that female participants would engage with the way society engages with information. While AI usage the tool. We hypothesize that the experiment will demonstrate is increasingly widespread, there is significant evidence that reduced engagement among women, suggesting that early male- indicates that there are large discrepancies in AI adoption rates dominated adoption may bias AI product development in ways between different demographic groups. This reality has far- that limit appeal and accessibility for underrepresented users and reaching implications for various domains ranging from workplace thus reinforce existing structural inequalities. This study lays the efficiency to information access. In this study, we investigate the groundwork for exploring how early user demographics influence downstream effects of gendered discrepancies in early AI adoption the design and adoption of generative AI tools. Future work could using an AI-powered news summarization tool. We survey 500 expand to include other demographic dimensions such as race, individualstoassessusagepreferencesandtrainapredictivemodel age, and education level, as well as examine longitudinal shifts in based on male users’ preferences. We then use this model to adoption behavior over time.
Abstract:
In the modern day, generative AI tools have rapidly reshaped evaluate the likelihood that female participants would engage with the way society engages with information. While AI usage the tool. We hypothesize that the experiment will demonstrate is increasingly widespread, there is significant evidence that reduced engagement among women, suggesting that early male- indicates that there are large discrepancies in AI adoption rates dominated adoption may bias AI product development in ways between different demographic groups. This reality has far- that limit appeal and accessibility for underrepresented users and reaching implications for various domains ranging from workplace thus reinforce existing structural inequalities. This study lays the efficiency to information access. In this study, we investigate the groundwork for exploring how early user demographics influence downstream effects of gendered discrepancies in early AI adoption the design and adoption of generative AI tools. Future work could using an AI-powered news summarization tool. We survey 500 expand to include other demographic dimensions such as race, individualstoassessusagepreferencesandtrainapredictivemodel age, and education level, as well as examine longitudinal shifts in based on male users’ preferences. We then use this model to adoption behavior over time.
Source:
Harvard / Harvard College | Mather House | Economics | 2028 / 2025
Topics:
adoption, tool, generative, early, demographic, user, evidence, summarization, engage, information, male, discrepancy
