Eye On A.I. #256 Stephen Schmidt: Inside Amazon’s AI-Powered Cybersecurity Strategy
May 22, 2025
Stephen Schmidt, Chief Security Officer at Amazon and former FBI counterintelligence expert, reveals Amazon's cutting-edge AI-powered cybersecurity strategies. He discusses the innovative MadPot honeypot network that tracks cyber threats in real time and the role of AI in alarm triage and threat detection. The conversation also delves into the security challenges of open-source vs. closed-source models and the implications of the $5M Nova Trusted AI Challenge to combat adversarial code. Schmidt emphasizes the critical importance of data privacy and secure AI deployment.
AI Snips
Chapters
Transcript
Episode notes
Rapid Honeypot Detection
- Amazon's MadPot honeypot network launches servers that appear vulnerable to attract attackers.
- Adversaries discover new honeypots in only 90 seconds and attempt exploits within three minutes.
AI Accelerates Attacker Tactics
- AI tools currently help adversaries mainly in generating convincing phishing and targeted emails.
- AI accelerates attackers' work but does not fundamentally invent new attack methods yet.
Choose AI Models Wisely
- Use open or closed-source AI models based on your business needs and risks.
- Build guardrails and ensure visibility to comply with regulatory requirements.