Automated decision making is subjected to systematic mistakes or systematic risk.
The sources of these risks are:
1) Genuine imbalances in the constituent populations
2) Thee way human beings train/inject bias in the algorithms
3) Limitations in the technology
As a manager you need to understand these limitations. These two articles describe use cases and opinion on the matter. They are very readable.
https://towardsdatascience.com/is-your-machine-learning-model-biased-94f9ee176b67
https://chatbotsmagazine.com/we-need-to-talk-about-biased-ai-algorithms-ead03bf964d5