Community Virtual Co-working 👩💻
Join here: meet.google.com/mkz-yyax-jnr
Come build and tinker in AI with other builders hardcore about creating new innovations with AI ✨.
Keep yourself accountable to build goals and get your AI questions answers. You will also get access to our platform!
Topic
Jailbreak Prompt Detection and Token-Based Attack Prevention
This session will be led by Avira, Security Engineer at Standard Chartered Bank
Are you passionate about AI safety and security? Don’t miss our upcoming virtual co-working session, where Avira, will lead an interactive demo on one of the most pressing topics in AI today: detecting and preventing jailbreak prompts and token-based attacks. She will share also how to detect and prevent prompt injection and token manipulation attacks targeting LLM-based systems — a growing concern for developers integrating AI into real-world applications.
What You’ll Learn
In this hands-on session, we’ll dive deep into:
Jailbreak Prompt Detection: Understand how attackers craft prompts to bypass AI safety mechanisms, and learn practical techniques to identify and mitigate these risks in real time.
Token-Based Attack Prevention: Explore how malicious actors exploit token manipulation to subvert language models, and discover robust, open-source methods to defend against these sophisticated attacks.
Open-Source Security Tools: Get step-by-step guidance on leveraging popular libraries and pre-trained models for effective security analysis in your own AI projects.
Don’t miss this opportunity to upskill, collaborate, and connect with fellow AI enthusiasts and security professionals!
Build Club AI Indonesia Team 🚀