Using LLMs to Find and Prioritize Vulnerabilities Is No Easy Task

Large language models used for vulnerability scanning have high false-positive rates, making it harder for AppSec professionals to prioritize vulnerabilities. This issue arises from the models' inability to consider the context of scans. As a result, professionals have to spend more time reviewing and validating scan results. This inefficiency highlights the need for more accurate and context-aware vulnerability detection methods. To mitigate this issue, consider implementing more advanced scanning tools or techniques.