The Malicious Use of AI Report

Summary

A report examining how AI technologies can be exploited by malicious actors across cybersecurity, physical security, and political domains. It analyzes near-term threats from AI misuse and offers recommendations for researchers, policymakers, and industry to mitigate these risks.

Key Points

• •Identifies three major threat domains: digital security (cyberattacks), physical security (autonomous weapons, drones), and political security (disinformation, surveillance).
• •Argues that AI lowers the cost and expertise required for malicious actors to conduct sophisticated attacks at scale.
• •Recommends dual-use research norms, responsible disclosure practices, and proactive engagement between AI researchers and security communities.
• •Highlights that AI can automate and scale social engineering, spear-phishing, and fake media generation for influence operations.
• •Calls for policymakers to anticipate and prepare for AI-enabled threats rather than react after harms occur.