Scholarlink

The malicious use of powerful generative AI tools took off just as quickly as their legitimate use. This chapter discusses the main vulnerabilities and threats of LLMs, and the main types of malicious generative AI use, from prompt injection and unauthorized code execution to deepfake videos that can potentially manipulate voters.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Offensive Generative AI: From Criminal LLMs to Deepfake-Based Deception

  • 章节

  • 发表日期：

    2025年

查看全文
  • Leslie F. Sikos

摘要

The malicious use of powerful generative AI tools took off just as quickly as their legitimate use. This chapter discusses the main vulnerabilities and threats of LLMs, and the main types of malicious generative AI use, from prompt injection and unauthorized code execution to deepfake videos that can potentially manipulate voters.