Understanding the New LLM Threat Landscape for Application Security
Large language models don’t just write code: they read it, reverse-engineer it, and hand attackers a roadmap through your application logic. What once took hours of manual effort now takes seconds.
Join Jscrambler host Katia Kupidonova and Jscrambler experts, Pedro Pereira, and Duarte Carneiro for a deep dive into how LLMs have fundamentally changed the threat landscape for client-side code, and what resilient protection looks like in response.
Alongside real-world examples, we share our perspective on the evolving challenges AI introduces for application security, and the defensive strategies that may become essential as LLM capabilities continue to advance.
You’ll walk away understanding:
- How LLM-powered attacks work step by step — and what makes them different from anything before
- The specific risks facing your industry, from entitlement logic to pricing manipulation to transaction flows
- What code resilience actually requires at runtime
- How Jscrambler has evolved its approach to stay ahead of LLM-based threats — and what’s coming in 2026
You’ll walk away understanding:
- Why obscurity is no longer a meaningful defense
- How LLM-powered attacks work step by step — and what makes them different from anything before
- What code resilience actually requires at runtime
- How Jscrambler has evolved its approach to stay ahead of LLM-based threats — and what’s coming in 2026
Meet the Speakers
Katia Kupidonova
Jscrambler, Product Marketing Manager
https://www.linkedin.com/in/katia-kupidonova/
Pedro Pereira
Jscrambler, AppSec/ML Researcher
https://www.linkedin.com/in/pedro-pereira-7abb53258/
Duarte Carneiro
Jscrambler, Product Manager
https://www.linkedin.com/in/duartecarneiro/
