AI-900 Question 65
Select 3You are developing an AI-powered recruitment tool to screen resumes and identify suitable candidates for job openings. To ensure inclusiveness in your AI solution, what steps should you take during the development process?
- A
Analyze the training data to identify and mitigate any biases present in the data.
- B
Ensure that the AI model prioritizes candidates with similar profiles to the current workforce to maintain consistency.
- C
Test the AI model on a diverse dataset that reflects different demographics and backgrounds.
- D
Provide transparency on how the AI makes decisions and allow human oversight of the recommendations.
- E
Exclude demographic information such as age, gender, or race from the training dataset to prevent bias.
Show answer and explanation
Correct answers: A, C, D
Explanation
Inclusiveness in AI solutions is achieved by addressing bias in the training data, testing the model on diverse datasets, and ensuring transparency and human oversight. These actions help prevent discrimination and promote fairness in AI-driven outcomes. Excluding demographic data alone is not sufficient to eliminate bias, and aligning AI recommendations to a homogeneous workforce undermines inclusiveness.
- A. Correct.
Analyzing and mitigating biases in the training data is a critical step to ensure the AI solution makes fair and inclusive decisions.
- B. Incorrect.
Prioritizing candidates similar to the current workforce risks reinforcing existing biases and does not promote inclusiveness.
- C. Correct.
Testing the AI model on a diverse dataset helps ensure that its performance is consistent across different groups and avoids unintentional discrimination.
- D. Correct.
Providing transparency and human oversight ensures accountability and trustworthiness in the AI’s decision-making process, which supports inclusiveness.
- E. Incorrect.
Simply excluding demographic information does not guarantee inclusiveness, as biases can still be present in other features or patterns within the data.