The road ahead: future trends, challenges, and opportunities in artificial intelligence-driven finance

Artificial intelligence (AI) is reshaping the world in every aspect. Financial industry is leading in the AI frontier for competitiveness and profitability. New AI techniques, such as deep learning and graph learning, are rapidly emerging and impacting the finance sector. In addition to the new techniques, there are huge challenges and opportunities in research and practice. This paper delivers a comprehensive analysis and exploration of AI in finance. It introduces the financial domain diversity, distinctly astonishing informatics and challenging issues related to the broad AI avenue in finance at domain, knowledge, modeling, infrastructure, method, and algorithm levels, and summarize horizon issues of overarching topics and booming trends, which can be chartered as a fiscal cosmos and analyzed as a galaxy. AI in finance has been broadly explored, with many methods proposed. Nevertheless, the quest for opportunities and the need for solutions remain. Broader, newer, and smarter financial AI methods, applications, and businesses are needed as high-impact, precedent-setting and high-benefits approaches to resolve the largely unsolved problems or challenges in the aspect of technique, knowledge, analytics, and infrastructure.

AI is not a new concept to finance. Employing mathematical models with the differences in interest exchange rates and security prices, arbitrage in options, futures, or bonds for profit has been widely used well before the advent of the term AI. Other than traditional Economics, Finance, and Political Science disciplines explored during the 20–30th centuries, plenty of finance/economics datasets have become open access. These intellectually leading datasets and unprecedented advances in information technology have accelerated vigorous interests in applying AI on wider, broader, generic, or bigger financial modeling, knowledge mining, and information prediction. The finance domain is a galactic world with tremendous astronomy (financial informatics), many celestial (financial) bodies (financial businesses), stars (operators/types of stocks), and even hundreds of thousands of satellites (customers/traders/entities). It is redefined, transformed, and appraised by deep distributed financial modeling through distributed diverse platforms or services involving multi-market/domain/channel/sector applications. Considerable analysis on cross-market financial data is still emerging with respect to the huge broadband and diversity of finance domain datasets.