Ligia Catherine Arias-Barrera (Externado de Colombia University) has posted Reconstructing Algorithmic Trading in the Age of Generative AI: Implications for Market Structure, Regulation, and Epistemology (Revista Emercatoria December 2025) on SSRN. Here is the abstract:
The purpose of this article is to analyse how generative artificial intelligence (Gen AI)– defined here as systems capable of producing novel outputs, such as text, code, or data, through machine learning– is transforming algorithmic trading. Gen AI enhances computational efficiency but also reconfigures the epistemological (theories of knowledge), institutional (market structures and actors), and regulatory (rules and oversight) foundations of financial markets. We contend that Gen AI ushers in a shift from price as a representation of economic fundamentals to price as a recursive output of machine-to-machine simulation. Drawing on lessons from financial machine learning, critical legal theory, political economy, and the sociology of finance, we study how generative models– such as transformer-based architectures (deep learning models initially developed for natural language processing) and synthetic data engines (tools generating artificial datasets)– reshape the logic of price discovery, liquidity provision, and risk assessment. Our analysis shows that Gen – AI intensifies market abstraction, undermines traditional regulatory assumptions, and consolidates informational power among a narrow elite of technologically sophisticated actors. It sheds light on epistemic opacity, feedback loops, and algorithmic reflexivity. These features compromise both market intelligibility and regulatory legitimacy. We also pay particular attention to how these dynamics exacerbate global inequalities. This happens through infrastructure asymmetries and imports regulatory templates that marginalize the global south.We conclude by urging the implementation of new regulations that prioritize clarity, accountability, and digital control. We suggest reforms such as public model registries, participatory audits, and open algorithmic infrastructure to restore democratic oversight in markets increasingly run by code. Our goal is to reposition algorithmic trading as a central arena for debates about value, governance, and economic reality in synthetic finance.