In HR, the same tool may safely help restructure a job description yet create significant risk if used to automatically reject candidates. The first question is not “can AI do it?” but “who is affected if it is wrong, and who can detect the error?”
Separate work into three groups. Lower-risk content support includes structuring JDs, summarizing policies, and drafting training questions. Controlled analysis includes coding anonymized feedback. High-impact decisions such as hiring, rejection, performance, discipline, and promotion require human authority.
Do not paste CVs, health data, compensation, disciplinary records, or identifying information into public AI tools. Practice with synthetic or anonymized data and use approved platforms for real data.
Your output is a one-sentence pilot scope: task, permitted data, AI output, reviewer, and decisions that remain human.
Try it now
- 1List five time-consuming HR tasks and the people affected
- 2Mark personal data, sensitive data, and relevant policies
- 3Rate risk based on the impact of an incorrect or biased output
- 4Choose a low- or medium-risk task with a qualified reviewer
- 5Complete the workbook's page-three canvas before testing a prompt
I want to assess this HR task: [description]. Do not process real data. Help me classify affected people, required data, prohibited data, a suitable AI role, reviewer, impact if wrong, and pilot stop criteria.
Practical AI for HR Workbook
Use the fill-in canvas on page three to complete this lesson with your own work task.
Download the workbook ↓Ready for the deeper practical lessons?
Return to the course overview and register. Aptus AI will send five deeper lessons over 14 days, each with a real-work assignment and quality check.
Join the free email course