AI Chatbots Make Web Attacks Worse — New Research Shows the Risk

Large language models are becoming core to web applications and browser agents, changing online interactions—but also reshaping old web threats and introducing new ones. According to the article, classic risks like cross-site scripting can be amplified through LLM-mediated interactions, while LLM-specific attacks like prompt injection can spread across web applications. Rather than treating web and LLM security separately, this survey offers a unified analysis of how LLMs amplify vulnerabilities across client-side, server-side, and pipeline layers, and evaluates defenses and their limits.

The article also looks at extending NIST and ISO/IEC AI security frameworks to LLM-enabled web environments. It identifies three unresolved challenges: adversarial natural-language instructions, autonomous agent security, and post-deployment security through continuous monitoring and adaptation. To address them, it proposes an LLM-aware monitoring and control framework that integrates semantic input validation, prompt integrity protection, output isolation, agent governance, and runtime monitoring—pointing toward future directions for secure AI-enabled web systems.