As the cliché goes: Garbage in, garbage out. That applies directly to AI-powered systems.
If the data going into the system is inaccurate or incomplete, then inferences from that data will be flawed.
It’s the same thing with data diversity. If an AI system only sees a limited subset of the environment, it ends up biased. It won’t make strong inferences about the rest of the system because the examples it’s learning from don’t represent the full picture.
This is especially true in incident management. When you can harvest incident…
