EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

3096. Generative AI Governance by Blockchain Technology

Invited abstract in session MC-4: New Trends in Generative Adversarial Networks and Deep Neural Networks , stream Recent Advancements in AI .

Monday, 12:30-14:00
Room: 1001 (building: 202)

Authors (first author is the speaker)

1. Swati Sachan
Management School, University of Liverpool
2. Vinícius Dezem
Knowledge Engineering and Management, Federal University of Santa Catarina
3. Dale Fickett
Robins School of Business

Abstract

Generative AI tools powered by Large Language Models (LLMs) have demonstrated advanced capabilities in generating and articulating coherent textual content closer to the practitioners in domains such as law and finance. Globally, firms have raised ethical concerns regarding LLM's ability to mimic human reasoning, accountability for erroneous outcomes, and the security and privacy of confidential data sent by the prompt of Generative AI. This research aims to find a balance between the responsible application of Generative AI and maintaining human oversight over the generated content by utilizing the inherent immutability and decentralization characteristics of blockchain technology. The proposed blockchain-based auditing system detects unauthorized alterations of data repositories containing ex-ante decisions by an AI decision-support system and automated textual explanations by Generative AI tools. The auditing algorithm compares the unique signature, known as Merkle roots of files stored off-chain (outside blockchain), with their immutable blockchain counterpart. Automated auditing by blockchain promotes the ethical use of AI technologies and minimizes the risk of discrepancies in attributing accountability for adverse decisions. A case study on pre-litigation tort liability legal cases is presented to demonstrate the practical application.

Keywords

Status: accepted

Back to the list of papers