Rachel is designing a new AI-powered application that will automatically match resumes with available job openings based on a wide variety of job applicant and job role characteristics. She plans to train the system using resumes of current employees in similar job positions who have received favorable reviews from their employers. However, Rachel recognizes there could be some bias represented in the current employment pool regarding employees' race, gender, marital status, and other characteristics. She wants to ensure that the AI system does not incorporate existing bias and make sure it evaluates potential employee matches based on their qualifications. At the same time, she hopes to give the AI system enough information so it can detect currently unknown patterns of what characteristics contribute to employee success in their job roles. Do some research online about what factors make for high quality training datasets. List three factors that will help Rachel choose the best training data for her AI application. What strategies can she use to help eliminate bias in her AI application?