Selective Response Strategies for GenAI

Boaz Taitler, Omer Ben-Porat

TL;DR: We show that GenAI and society can both benefit from selective response of GenAI.

Abstract: The rise of Generative AI (GenAI) has significantly impacted human-based forums like Stack Overflow, which are essential for generating high-quality data. This creates a negative feedback loop, hindering the development of GenAI systems, which rely on such data to provide accurate responses. In this paper, we provide a possible remedy: A novel strategy we call selective response. Selective response implies that GenAI could strategically provide inaccurate (or conservative) responses to queries involving emerging topics and novel technologies, thereby driving users to use human-based forums. We show that selective response can potentially have a compounding effect on the data generation process, increasing both GenAI's revenue and user welfare in the long term. From an algorithmic perspective, we propose an approximately optimal approach to maximize GenAI's revenue under social welfare constraints. From a regulatory perspective, we derive sufficient and necessary conditions for selective response to improve welfare.

Lay Summary: When people use LLM, they don't generate data in public Q&A forums. Yet, such human-generated data is vital for training future LLMs. We propose a way, which we call selective response, where LLMs strategically provide partial or conservative responses, encouraging users to seek answers on human-driven forums occasionally. We build a theoretical model to test this strategy, identifying scenarios where it benefits both users and the LLM itself, and propose practical algorithms for real-world implementation.

Primary Area: Theory->Game Theory

Keywords: game theory, generative ai

Submission Number: 6426