research papers

  • Free Speech & the Fake News Problem

    Authors: Marshall W. Van Alstyne
    Abstract: This paper explores the intersection of free speech and the rise of fake news. We examine regulatory solutions and propose new frameworks for addressing this issue in online media.

  • Secure Account Recovery for a Privacy-Preserving Web Service

    Authors: Ryan Little, Lucy Qin, and Mayank Varia
    Abstract: This paper presents a cryptographic protocol for secure account recovery in a privacy-preserving web service, ensuring that users can recover their accounts without revealing personal data to the provider. The proposed system leverages standardized cryptographic techniques, including an oblivious pseudorandom function, to enable secure email-based validation while preventing unauthorized access.

  • Improving the Governance of Digital Platforms with Interactive Marketplace Experiments

    International Conference on Computational Social Science, 2024

    Authors: Swapneel Mehta, Sverre Wiedswang, Zoey Yang, Nina Mazar, Marshall van Alstyne
    Abstract: An empirical experiment discussed in our extended abstract for IC2S2, 2024 where we present our findings from the deployment of self-certification, or truth warrants in a competitive two-sided e-commerce marketplace.

  • Market Design Interventions for Safer Agentic AI

    Authors: Swapneel Mehta, Aaron Nichols, Abhishek Shah, Pratyay Banerjee, Nina Mazar, Marshall van Alstyne
    Abstract: In this paper, we run an empirical study with human buyers purchasing products based on ads generated by large language model ‘agents’ to find an increase in misleading advertising and consumer harm. Using our novel economic intervention we demonstrate how to curtail deception by dishonest agentic advertisers while improving the profits of honest ones using truth warrants.

talks

  • Improving the Governance of Online Marketplaces using Warrants

    Author: Swapneel Mehta
    Venue: Stanford Trust and Safety Research Conference (2024)

  • Market Design Interventions for Safer Agentic AI

    Presenter: Aaron Nichols
    Venue: Yale AI/ML Conference (2024)
    Venue: Columbia MarkTech Conference (2024)

  • Certifiably True: The Impact of Self Certification on Misinformation

    Author: Aaron Nichols
    Venues:

    • Platform Strategy Research Symposium (2024)
    • International Conference on Computational Social Science (2024)
    • Freedom of Expression Scholars Conference (2024)

press