The U.S. Defense Advanced Research Projects Agency (DARPA) is exploring how AI and automated reasoning could help identify vulnerabilities in the design of electronic markets before adversaries exploit them, according to a November 8 report by Business Insider. 

Its Intrinsic Market Resilience (IMR) program targets systems supporting financial markets, energy trading and cloud-computing resources.

The initiative, announced in a DARPA solicitation, wants to identify weaknesses embedded in the rules and mechanisms that govern digital markets, which could be exploited to disrupt market operations or create wider economic instability.

DARPA’s proposed approach combines large language models with formal mathematical analysis and automated reasoning, according to the report. The agency wants researchers to use AI to interpret regulatory filings, technical specifications and other documents, then convert that information into computer models of how electronic markets operate.

Those models could be tested against defined resilience standards to identify conditions under which a market might fail or behave unpredictably. Researchers would then develop examples of potential attacks and recommend changes to the underlying market design.

Matthew Wilding, the program’s manager, told Business Insider that advances in automated reasoning and AI are creating new ways to defend complex systems, even as the same technologies could also be used to attack them.

Why Market Design Is a National Security Concern

The initiative reflects a broader recognition that vulnerabilities in digital infrastructure do not always originate in software flaws or stolen credentials. Weaknesses in the rules governing how systems interact can also be exploited.

The California electricity crisis in the early 2000s, for example, demonstrated how market mechanisms could be manipulated to produce severe economic consequences. The 2010 flash crash also exposed how trading behavior and market structures can interact in unexpected ways.

DARPA wants its researchers to identify comparable weaknesses before malicious actors can take advantage of them. The aim is to help designers and regulators understand how a system could be manipulated and make changes that improve its resilience.

The program therefore connects AI research with the security of economic infrastructure, including systems that support financial transactions, energy markets and digital services.

Implications for Data and Technology Leaders

For chief data officers (CDOs), the project offers an example of how AI can be applied to test complex systems and identify risks that conventional monitoring may overlook. 

It also highlights the importance of data quality and access to technical documentation when building reliable models of interconnected environments.

The work could eventually inform how financial institutions, infrastructure operators and regulators assess the resilience of their own digital systems. AI-assisted modelling may help them examine scenarios that would be difficult, costly or disruptive to test in live environments.