- Did its training data set mimic a diverse environment?
- Does the AI use demographic data (race, gender, etc) or inferring data (university attended, name, etc) to suggest matches or candidates?
- Is the AI monitored or regularly audited by a neutral third party?



