Original title: Cyber-Financial Contagion: Modeling the Propagation of an AI Vendor Compromise Through the Banking System
Alex Leytes · September 2026
Many banks rely on the same few artificial intelligence companies to screen for fraud, approve loans, and analyze customers. If an attacker hacks just one of these shared vendors, the damage can spread quickly. Researchers wanted to see if such an attack could cause a full financial crisis. They built a computer simulation that links AI companies, banks, bank-to-bank loans, and customer accounts. In this model, an infection passes through computer systems and money connections like a spreading virus. The team found that a single vendor hack can trigger massive financial losses. The total damage depends heavily on how slowly banks fix the security flaws. Finally, they built an early-warning system that spots high-risk AI vendors before an attack causes wide harm. This shows that relying on a few vendors threatens entire financial systems.
Imagine a neighborhood where several houses share a single water pipe. If that pipe gets contaminated, every house gets dirty water. Now imagine that the contamination can also spread from house to house through shared plumbing. In this paper, the shared AI vendors are like the main water pipe, and the banks are the houses. A compromise in the vendor contaminates all connected banks, and then the problem can spread further through bank-to-bank connections.
Regulators and banks now rely on a handful of AI vendors for critical operations. If one vendor is compromised, the damage can spread through the banking system like a classic financial crisis. This paper gives supervisors a concrete tool to model and predict such cascades, helping prevent systemic failures before they happen.
The work uses synthetic data, not real banking data, so real-world applicability is untested. The early-warning model has moderate accuracy (AUROC 0.82) and may miss some cascades. The model assumes certain network structures and does not account for all possible contagion mechanisms.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Cyber-Financial Contagion: Modeling the Propagation of an AI Vendor Compromise Through the Banking System — Alex Leytes
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.