Once a detection model is selected you will be able to pick from a variety of AI/ML algorithms such as Naive bayes, Support vector machine, or a neural network. Each detection model has been trained using varies AI/ML algorithms.
Penetration testers, red teams, bug bounty hunters, and anyone else in the offensive security security space has to deal with bypassing detection mechanisms.
Scan your phishing domains & emails, malware, and exploits to see if they are detected. Since so many vendors are incorporating AI/ML into their cyber security products its a good idea to make sure your payloads can bypass these techniques.