Inside DeepSeek’s Security Flaws

Artificial intelligence is advancing fast, but with innovation comes risk. In this episode of Threat Vector, host David Moulton sits down with Sam Rubin, SVP of Consulting and Threat Intelligence at Unit 42, and Kyle Wilhoit, Director of Threat Research, to explore the vulnerabilities of DeepSeek, a new large language model. They dive into AI jailbreaking techniques like Bad Likert Judge and Deceptive Delight, uncovering how attackers can manipulate models to generate harmful content. The discussion highlights security risks, data integrity concerns, and why businesses must carefully evaluate AI tools before adopting them. Tune in for expert insights on the evolving cybersecurity landscape and what organizations can do to stay protected. Join the conversation on our social media channels: Website: http://www.paloaltonetworks.com Threat Research: ⁠⁠⁠⁠https://unit42.paloaltonetworks.com/⁠⁠⁠⁠ Facebook: ⁠⁠⁠⁠https://www.facebook.com/LifeatPaloAltoNetworks/⁠⁠⁠⁠ LinkedIn: ⁠⁠⁠⁠https://www.linkedin.com/company/palo-alto-networks/ YouTube: ⁠⁠⁠⁠@paloaltonetworks Twitter: ⁠⁠⁠⁠https://twitter.com/PaloAltoNtwks⁠⁠⁠⁠ About Threat Vector Threat Vector, Palo Alto Networks podcast, is your premier destination for security thought leadership. Join us as we explore pressing cybersecurity threats, robust protection strategies, and the latest industry trends. The podcast features in-depth discussions with industry leaders, Palo Alto Networks experts, and customers, providing crucial insights for security decision-makers. Whether you're looking to stay ahead of the curve with innovative solutions or understand the evolving cybersecurity landscape, Threat Vector equips you with the knowledge needed to safeguard your organization. Palo Alto Networks Palo Alto Networks enables your team to prevent successful cyberattacks with an automated approach that delivers consistent security across the cloud, network, and mobile. ⁠http://paloaltonetworks.com⁠