DDoD: Dual Denial of Decision Attacks on Human-AI Teams

TL;DR

This paper introduces DDoD attacks that target both human and AI computational resources to disrupt decision-making in human-AI collaborative systems, highlighting new vulnerabilities and potential risks.

Contribution

It proposes a novel dual-resource denial attack framework against human-AI teams, expanding the scope of adversarial threats beyond traditional AI model attacks.

Findings

DDoD attacks can deplete both human and computational resources.

Such attacks significantly impair decision-making in collaborative systems.

Potential risk scenarios demonstrate the severity of these attacks.

Abstract

Artificial Intelligence (AI) systems have been increasingly used to make decision-making processes faster, more accurate, and more efficient. However, such systems are also at constant risk of being attacked. While the majority of attacks targeting AI-based applications aim to manipulate classifiers or training data and alter the output of an AI model, recently proposed Sponge Attacks against AI models aim to impede the classifier's execution by consuming substantial resources. In this work, we propose \textit{Dual Denial of Decision (DDoD) attacks against collaborative Human-AI teams}. We discuss how such attacks aim to deplete \textit{both computational and human} resources, and significantly impair decision-making capabilities. We describe DDoD on human and computational resources and present potential risk scenarios in a series of exemplary domains.