Note dell’episodio
In this episode of the Ransomware Rewind podcast, host Joe Erle (@joe_erle) interviews John Bruggeman, Chief Information Security Officer (CISO) at CBTS and OnX, on emerging cybersecurity threats like AI model poisoning and prompt injection attacks. With over 25 years of experience in cybersecurity, John explains how unsanitized inputs and as few as 250 malicious data points can cause "brain rot" or model decay in large language models (LLMs), resulting in unreliable outputs, hidden backdoors, and long-term AI vulnerabilities.
John explains real-world AI attack vectors, including tool…
Sito web: 10d31dcf-2599-4c6a-818b-c4547298bd9c.libsyn.com

