CYAIMar 22, 2024

A Technological Perspective on Misuse of Available AI

arXiv:2403.15325v110 citationsh-index: 4
Originality Synthesis-oriented
AI Analysis

It addresses the security risks of AI misuse for national and international stakeholders, highlighting an incremental analysis based on current technologies.

The paper examines how existing, openly available AI technologies can be maliciously combined to create autonomous weapon systems, posing threats to political, digital, and physical security, and proposes control measures to prevent such misuse.

Potential malicious misuse of civilian artificial intelligence (AI) poses serious threats to security on a national and international level. Besides defining autonomous systems from a technological viewpoint and explaining how AI development is characterized, we show how already existing and openly available AI technology could be misused. To underline this, we developed three exemplary use cases of potentially misused AI that threaten political, digital and physical security. The use cases can be built from existing AI technologies and components from academia, the private sector and the developer-community. This shows how freely available AI can be combined into autonomous weapon systems. Based on the use cases, we deduce points of control and further measures to prevent the potential threat through misused AI. Further, we promote the consideration of malicious misuse of civilian AI systems in the discussion on autonomous weapon systems (AWS).

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes