TechStream, a rising tech start-up, developed an AI-powered chatbot for its clients’ websites.

The root cause of this incident is unvalidated input poisoning the AI training process, resulting in malicious output. ECIH web application security principles emphasize that input validation is the primary control against injection and manipulation attacks.

Option B is correct because strict validation ensures that malicious or untrusted data cannot influence training or response generation. This addresses the vulnerability at its source.

Option A adds friction but does not stop poisoning. Option C is disruptive and not efficient. Option D limits functionality without fixing the underlying issue.

ECIH stresses fixing vulnerabilities at the root rather than applying superficial controls, making Option B correct.