This report surveys the landscape of potential security threats from malicious uses of AI, and proposes ways to better forecast, prevent, and mitigate these threats. After analyzing the ways in which AI may influence the threat landscape in the digital, physical, and political domains, we make four high-level recommendations for AI researchers and other stakeholders. We also suggest several promising areas for further research that could expand the portfolio of defenses, or make attacks less effective or harder to execute. Finally, we discuss, but do not conclusively resolve, the long-term equilibrium of attackers and defenders.

翻译：本报告系统梳理了人工智能恶意使用可能带来的安全威胁态势，并提出了改进威胁预测、防范与缓解的路径。在分析人工智能可能影响数字、物理及政治领域威胁格局的作用机制后，我们向人工智能研究者及其他利益相关方提出四项高层建议。同时指出若干具有前景的后续研究方向，这些方向有望拓展防御手段体系，或降低攻击效能与可实施性。最后，我们探讨了攻击者与防御者之间的长期均衡状态（但未给出确定性结论）。