2026 ASEE Annual Conference & Exposition

Businesses have problems with zero-day vulnerabilities when releasing products. The problem roots from the fact that zero-day vulnerabilities take a lot of time and people to discover. While businesses try to make their products as secure as possible, sometimes there are vulnerabilities that they are unable to find and they end up having to release the product before they can find everything. This is where white hats and ethical hackers come in to aid in finding the vulnerabilities that the companies could not find in their product upon release. While there are bounties set on finding these, there are also people that will find these vulnerabilities and sell information about them to others that may have malicious intentions with such information. In order to avoid these situations, there could be AI models that are trained to specialize in finding oversights in products that can aid and accelerate the discovery of zero-day vulnerabilities by the ethical hackers or companies. Due to the power of something like this, not everyone should be able to access this tool. To keep it from being used by bad actors and for its intended use of supporting white hats, there should be an Open Weights model in effect to give only what the training has resulted in rather than what data was used to train the model.
This is my original work and is not being reviewed anywhere else.