Using Code Inspection, Code Modification, and Machine Learning to prevent SQL Injection

Modern day databases store invaluable information about everyone. This information is assumed to be safe, secure, and confidential. However, as technology has become more widespread, more people are able to abuse and exploit this information for personal gain. While the ideal method to combat this issue is the enhanced education of developers, that still leaves a large amount of time where this information is insecure. This thesis outlines two potential solutions to the problem that SQL Injection presents in the context of databases. The first modifies an existing code base to use safe prepared statements rather than unsafe standard queries. The second is a neural network application that sits between the user-facing part of a web application and the application itself. The neural network is designed to analyze data being submitted by a user and detect attempts at SQL injection.