@Sysengineer honestly I always say it comes down to the individuals monitoring and training anything AI/ML. Like anything else it’s subject to bias, false positives, etc. if you’re relying on it as an org, it should be used as a “in addition to” with other tools.

As long as it’s getting fed accurate and correct data via the humans - then it’s always a great back pocket tool. Even if just to validate results from another tool.