Review of Gen AI in Fixed Income Markets: Trading, Modeling and Risk Management

Open Webpage

Satyadhar Joshi

Published: 20 Oct 2025, Last Modified: 22 Nov 2025figshareEveryoneRevisionsCC BY-SA 4.0
Abstract: <p dir="ltr">Abstract: This paper presents a systematic review of generative artificial intelligence (AI) applications in fixed income markets, synthesizing insights from key studies published mostly between 2024 and 2025. The analysis spans the latest developments in AI-driven analytics, trading strategies, risk management techniques, and the evolution of investment approaches within this crucial sector of finance. The review covers advancements in interest rate yield curve modeling, algorithmic trading, credit and liquidity risk assessment, and structured product valuation. Special attention is given to the emergence of large language models (LLMs), such as BondGPT and ChatGPT, and their integration into bond analytics, scenario generation, and regulatory workflows. These technologies are shown to enhance efficiency in bond analytics, trade documentation, liquidity analysis, and credit research, while also introducing new challenges related to model risk, data integrity, and compliance. Our findings highlight three core areas of innovation: (1) improved forecasting accuracy in interest rate models using hybrid AI architectures, (2) proposals for enhanced efficiency and automation in asset-backed securities (ABS) and mortgage-backed securities (MBS) pricing, and (3) explainable AI frameworks for compliance in risk-sensitive environments. Despite notable efficiency gains—ranging from 22% to 40% across key market functions—challenges persist around interpretability, data quality, and regulatory acceptance. Here, AI-driven valuation frameworks, scenario generation, and risk modeling innovations are accelerating the analysis and management of complex securities. Both the theoretical progress and practical impact of these models are explored in the paper, showing their abilities in explaining market patterns and their difficulties in meeting regulations. The study combines perspectives from academia, regulators and industry players to fully assess how AI is impacting the world of fixed income and offers suggestions for future improvements and best approaches.</p>