Team
Principal Investigators
-
Jon Kleinberg is leading the design of the auditing component of the infrastructure, developing protocols for auditing the behavior of AI agents in online interactions and for assessing whether mediators and assistants comply with their stated roles and specifications.
WebsiteJon Kleinberg
Tisch University Professor of Computer Science at Cornell University
-
Cristian Danescu-Niculescu-Mizil is leading the project’s experimental and evaluation infrastructure, including the development of the benchmarking platform, simulation platform, and the evaluation toolkit. He is also responsible for the open-source tools enabling communities to design and implement AI mediators and assistants for supporting civic discourse.
WebsiteCristian Danescu-Niculescu-Mizil
Professor of Information Science at Cornell University
-
Natalie Bazarova is leading the development of community-tunable specifications for AI mediators and agents, focusing on how communities can define the norms, goals, and constraints that AI systems should follow when intervening in online conversations.
WebsiteNatalie Bazarova
Professor of Communication and Information Science at Cornell University
-
David Rand is leading the evaluation of the human impact of the Auditing Layer, designing experiments to test how people respond to AI mediation with and without visible auditing, certification, or verification mechanisms.
WebsiteDavid Rand
Professor of Information Science and Marketing and Management Communications at Cornell University
-
Robert Kleinberg is developing cryptographic-inspired approaches for auditing black-box AI agents, focusing on methods that can support verification, accountability, and trust without requiring access to the internal workings of the agents.
WebsiteRobert Kleinberg
Professor of Computer Science at Cornell University
-
Mor Naaman is leading the conceptualization and design of a general-purpose approach to trust and accountability of online sources, building and expanding on the Auditing Layer. He is leading the development of a sample application for highlighting trustworthy accounts and sources in online discussion environments.
WebsiteMor Naaman
Associate Dean and Don and Mibs Follett Professor of Information Science at Cornell Tech
Members
-
Cai Yang is working on the experimental workflow of ConvoArena, including the design and deployment of human studies involving AI mediators and agent participants that support different evaluations of conversational AI systems. He is also developing the AI mediator toolkit, which enables users to build and evaluate AI mediators for online conversations. Its core capabilities include reusable and configurable mediator templates, testing across different experimental settings, simulations and evaluations with AI agents.
LinkedInCai Yang
Research Scholar at National University of Singapore
-
Christina Bai is working on agent development, making AI agents more realistic through a multi-stage approach. She also integrated SimulationFactory into ConvoArena. In addition, she will develop the AI Agent Toolkit, which will enable users to build and evaluate AI agents.
LinkedInChristina Bai
Undergraduate at Cornell University
-
Claire Okamoto is developing the ConvoKit AI-Mediation Extension, including the conversational analysis pipeline that integrates ConvoArena and Simulation Factory outputs with ConvoKit.
LinkedInClaire Okamoto
Undergraduate at Cornell University
-
Jocelyn Zhu is working on the Public Auditing Protocol and on cryptographically-inspired approaches to auditing, focusing on how to formalize the information that AI mediators should expose for inspection, how to represent mediator roles and specifications in auditable form, and how to define procedures for detecting bias, inconsistency, manipulation, or other failures in black-box AI agents.
WebsiteJocelyn Zhu
PhD Student in Public Policy at Harvard Kennedy School
-
Juhyun Sung is working working on the Simulation Factory, especially on enabling fine-grained control of AI-mediated communication simulations.
LinkedInJuhyun Sung
Undergraduate at Cornell University
-
Laerdon Kim is working on the Simulation Factory, especially on enabling fine-grained control of AI-mediated communication simulations. In addition, he is developing and deploying ConvoCompass, the front-end for AI mediation and assistance in real online communities (currently focusing on Wikipedia and Reddit).
WebsiteLaerdon Kim
Undergraduate at Cornell University
-
Mandy Luo is working on the Community Specification Toolkit, helping design community-tunable specifications for AI mediators and agents. In addition, she is developing the ConvoKit AI-Mediation Extension, including the conversational analysis pipeline that integrates ConvoArena and Simulation Factory outputs with ConvoKit.
LinkedInMandy Manxi Luo
PhD Student in Communications at Cornell University
-
Marcelo Sartori is working on the Public Auditing Protocol and on cryptographically-inspired approaches to auditing, focusing on how to formalize the information that AI mediators should expose for inspection, how to represent mediator roles and specifications in auditable form, and how to define procedures for detecting bias, inconsistency, manipulation, or other failures in black-box AI agents.
LinkedInMarcelo Sartori Locatelli
PhD Student in Computer Science at Cornell University
-
Reed Orchinik is an Assistant Professor of Social and Decision Sciences at Carnegie Mellon University. He is helping with the human-evaluation workflow of ConvoArena, leading the study design of benchmarking scenarios.
WebsiteReed Orchinik
Assistant Professor in Department of Social and Decision Sciences at Carnegie Mellon University
-
Sophia Liu is working on the ConvoArena backend, implementing the capacity to run parallel live AI-mediated conversations with human participants. She is also developing the AI mediator toolkit, which enables users to build and evaluate AI mediators for online conversations. Its core capabilities include reusable and configurable mediator templates, testing across different experimental settings, simulations and evaluations with AI agents.
WebsiteSophia Liu
Undergraduate at Cornell University
-
Vivian Nguyen is providing scientific coordination across all project components.
WebsiteVivian Nguyen
PhD Student in Computer Science at Cornell University
