When AI Turns Criminal: Deepfakes, Voice-Cloning & LLM Malware

Bishop Fox breaks down how deepfakes and voice clones are built — data collection, training, inference — cites a $25M Hong Kong fraud, and outlines defenses against AI-driven social engineering.

Modern AI voice cloning needs only a minute or two of audio to produce realistic replicas. The article examines real-time impersonation risks, NetSPI's cloner testing tool, and mitigations.

Why patches fail: of 8,100 fixes studied, 15.5% needed follow-up repairs; case studies like SharePoint ToolShell show bypasses within weeks — and what defenders should re-check.

Revisits David Bianco's Pyramid of Pain and how security teams can apply it in practice, prioritizing indicators that cause attackers real pain to guide detection investments.

Bishop Fox Wrapped: the year's most-revisited research, sessions, and posts, covering social engineering, vulnerability research, mobile testing, deepfakes, IoT hacking, and LLM pen testing.

Dark Reading panel recap with Bishop Fox's Rob Ragan: AI lowers the bar for phishing, deepfakes and agentic prompt injection, but also powers defense — practical mitigations matter more than hype.