Weaponizing Intelligence: Understanding LLM-Driven Malware and Zero-Day Threats by Samita Bai

Описание: This webinar is part of the Cyber Pulse by CIC series hosted by Sumit Kundu.

Abstract: As generative AI reshapes the digital landscape, its misuse in cybersecurity threats is becoming alarmingly evident. This webinar focuses on the offensive capabilities of Large Language Models (LLMs), particularly in the context of malware generation and zero-day attack facilitation. We explore how LLMs—when manipulated through techniques like prompt injection, character play, and contextual redirection—can autonomously produce malicious code, craft obfuscated payloads, and even simulate exploit patterns that mimic zero-day vulnerabilities. With minimal technical expertise, attackers can now harness these models to lower the barrier for sophisticated, evasive threats. Join us as we examine these emerging risks, highlight real-world examples, and discuss the urgent need for resilient AI safeguards and proactive defense mechanisms.

Biography: Dr. Samita Bai, Ph.D. is currently working as a Postdoctoral Fellow at the Canadian Institute for Cybersecurity at the University of New Brunswick. Dr. Bai holds a Ph.D. in Computer Science from the Institute of Business Administration, Karachi, Pakistan. She has been an Assistant Professor at Salim Habib University, Karachi, Pakistan, with extensive experience in AI, Machine Learning, Cybersecurity, and Data Science. Her research focuses on Large Language Models (LLMs), Explainable AI (XAI), AI ethics, and Information Retrieval.

Dr. Bai is a Senior IEEE member and has actively contributed to Women in Engineering (WIE), previously serving as the Chair of the IEEE WIE Affinity Group, Karachi Section, Pakistan, and currently serving as the Treasurer of the IEEE WIE Affinity Group, New Brunswick Section, Canada.
-------------------------------
To learn more about the Canadian Institute for Cybersecurity watch,    • Canadian Institute for Cybersecurity  

#CybersecurityAwareness #Cybersecurity #AIThreats #LLM #ZeroDay #GenerativeAI #UNB #CIC #Webinar #WomenInSTEM #AIethics

Stay connected with us!
Twitter:   / cic_unb  
Facebook: https://fb.me/cicunbca
LinkedIn:   / canadian_institute_cybersecurity  
Blog: https://cyberdailyreport.com/blog
Website: https://www.unb.ca/cic/

Canadian Institute for Cybersecurity
University of New Brunswick
46 Dineen Drive, Fredericton, NB E3B 9W4

0:00 Opening Remarks
1:51 Introduction
2:18 Agenda
2:29 About Me
4:31 Rise of Generative AI
5:44 Large Language Models
6:31 Growth of LLMs
8:52 LLMs in Cybersecurity
10:23 AI-as-a-Service Models
11:45 Known AI-as-Service Models
12:37 Threat Alert!
13:14 Threat Landscape
15:34 Likelihood of Threat
16:00 New Research: Palo Alto Networks
17:04 Threat Categories Enabled by LLMs
18:51 Why Malware Generation with LLMs Matters?
20:19 Recent Research Initiatives
20:32 Obstacles in Generating Malware Using LLMs
23:13 Malware Generation Techniques
28:08 Jailbreaking
32:16 Types of Prompts
33:40 Common Jailbreaking Techniques
38:12 Jailbreaking Research for Malware Generation
40:00 Zero-Day Attacks
42:44 Rise of Zero-Day Exploits
43:24 LLMs for Zero-Day Exploitation
46:54 Ethical & Security Implications
48:09 Cyber Resilience in the Age of LLMs
50:52 Building Resilience into AI Systems
51:57 Conclusion
53:36 Closing Remarks and Q&A