Artificial Intelligence in Financial Decision-Making

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Provides a comprehensive taxonomy of AI methods in finance+ESG and identifies research gaps in integrated frameworks.

🏢実務担当者:Highlights how AI can be leveraged for ESG rating prediction and disclosure analysis, aiding sustainability reporting and investment decisions.

🏛政策担当者:Suggests regulatory attention to AI-driven ESG evaluation and potential risks such as interpretability and regime instability.

Artificial intelligence has become a major methodological force in financial decision-making, but the literature remains fragmented across at least three partially connected domains: financial time-series forecasting, portfolio construction, and firm-level sustainability analysis. This review argues that these domains should be interpreted as parts of a broader decision architecture in which algorithms extract signals from noisy data, transform those signals into investment or financing choices, and then evaluate outcomes under multiple objectives that increasingly include environmental, social, and governance criteria. The review first synthesizes the evolution of forecasting methods from classical econometric models to recurrent neural networks, transformers, and hybrid architectures. It then examines how predictive outputs are translated into allocation rules, with emphasis on mean–variance optimization, shrinkage-based risk estimation, risk parity, hierarchical allocation, and reinforcement-learning-based dynamic rebalancing. The third substantive line concerns corporate finance and sustainable finance, where AI is used not only to predict ESG ratings and financial constraints but also to identify firm heterogeneity, financing frictions, and disclosure-based signals. Across these streams, the article compares predictive and explanatory models, clarifies the role of structured, textual, and alternative data, and evaluates major methodological risks including overfitting, regime instability, interpretability deficits, and institutional dependence. The central conclusion is that the next stage of research should not treat forecasting, allocation, and ESG-related corporate finance as separate literatures. Instead, future work should build integrated frameworks in which market prediction, portfolio design, and firm-level sustainable finance analysis are jointly modeled under explicit assumptions about data quality, decision frequency, and accountability.

• semanticscholar https://jandoopress.com/journal/jgtss/article/download/250/241first seen 2026-07-18 08:31:02