![Join Our Paper Club Event with Vijay Keswani on "On The Stability of Moral Preferences: A Problem with Computational Elicitation Methods" [AI Tinkerers - Paper Club]](https://sloppy-joe-app.imgix.net/blog_images/paper-club-page-logo-png-Ozj_.png?w=300&fit=max&auto=format%2Ccompress&q=75){kind=logo}
Join Our Paper Club Event with Vijay Keswani on "On The Stability of Moral Preferences: A Problem with Computational Elicitation Methods"
Event Ended
This event has already taken place.
Join Our Paper Club Event Series! Meet with Vijay Keswani, AI Researcher, Duke University and discuss the problems of alignment and ethics in AI systems.
Don’t miss this unique opportunity: Hear directly from the researcher & join a live Q&A!
☝️ Register Above for this Live Virtual Meeting with the Researcher! ☝️
| Info | Details |
|---|---|
| Event | Paper Club with Vijay Keswani on “On The Stability of Moral Preferences: A Problem with Computational Elicitation Methods” |
| Date & Time | November 5, 2024, 12:00 PM EST |
| Presenter | Vijay Keswani, AI Researcher, Duke University |
| Research Paper | 📄 On The Stability of Moral Preferences: A Problem with Computational Elicitation Methods |
| Audio Version |
By Paper2Audio |
Meet the Researcher
Meet Vijay Keswani, AI researcher focused on algorithmic fairness, tech ethics, and community-oriented AI as a Postdoctoral Associate at Duke University. His research on the stability of moral preferences in AI systems investigates how moral preferences can change over time and the implications for AI tools that rely on these preferences.
Paper abstract:
Read the Paper: On The Stability of Moral Preferences: A Problem with Computational Elicitation Methods
This paper examines the stability of moral preferences in participants, which is crucial for ethical AI frameworks that rely on preference elicitation to reflect stakeholders’ values.
Here’s a problem: Typically, surveys assessing moral judgments are administered once per participant, assuming responses are stable over time. However, little is known about how consistent these responses truly are. To investigate this, participants repeatedly answered the same moral questions—specifically about kidney allocation decisions—across ten sessions in two weeks, with only question order varying.
Findings showed that participants changed their responses 10-18% of the time, especially in controversial scenarios, with changes linked to response time and decision difficulty. This instability suggests that moral preference elicitation could misalign AI decision-making with stakeholders’ actual values, raising methodological concerns for AI ethics and value alignment.
What is Paper Club?
Paper Club is a virtual event series brought to you by the Human Feedback Foundation in collaboration with AI Tinkerers, featuring authors of cutting-edge AI and machine learning papers. These online meetups allow attendees to hear about groundbreaking research directly from the authors, participate in live Q&A sessions, and engage in discussions. Open to all, Paper Club offers a regular opportunity to learn and interact with leaders in the rapidly evolving field of artificial intelligence.