DARPA wants AI to find hidden flaws in digital markets that adversaries could exploit to wreak economic havoc

Business Insider
Read full story· 6 min
DARPA is developing AI tools to detect hidden vulnerabilities in digital markets that could be exploited by adversaries to cause economic disruption. This initiative aims to enhance US defense and regulatory capabilities against cyber threats targeting economic systems.

What happened

The Defense Advanced Research Projects Agency (DARPA) has launched a new initiative to develop artificial intelligence tools that can identify hidden weaknesses in the rules governing US digital markets. The program, called Intrinsic Market Resilience (IMR), aims to protect electronic markets such as those for stocks, bonds, energy, and cloud computing from being exploited by adversaries. DARPA is seeking to use large language models to analyze regulatory filings and technical documents to model market operations and spot vulnerabilities. The goal is to detect and address flaws before they can be used to cause economic disruption.

Key details

  • The Intrinsic Market Resilience (IMR) program is led by DARPA.
  • The focus is on electronic markets, including stocks, bonds, energy, and cloud computing resources.
  • DARPA wants to use large language models to turn regulatory and technical documents into computer models of market behavior.
  • The program is not targeting profit-driven manipulation but rather adversaries aiming for economic disruption.
  • Historical incidents cited include the early 2000s California energy crisis and the 2010 'flash crash'.

Why it matters

US defense agencies, federal regulators, and market operators could gain new tools to detect and fix vulnerabilities in digital markets before they are exploited. This could reduce the risk of large-scale economic disruption caused by adversaries targeting market design flaws. The initiative may require market participants and regulators to adapt to new standards and processes for market resilience.


More in Cybersecurity