Chapter 10
Conclusions and roadmap
From technological promises to professional practice that is accountable and open to evaluation.
10.1 Synthesis of the argument
Artificial intelligence can expand analytical capabilities and reduce certain administrative tasks. Its value depends on process design, data quality and the legitimacy of its uses. Benefits must be demonstrated in a specific context rather than inferred from a technological label.
People management requires integrating efficiency, fairness, privacy and participation. No single metric summarizes these dimensions. Decisions must acknowledge their tensions and explain how the perspectives of those bearing the consequences were considered.
10.2 A staged roadmap
The first stage identifies a need and describes the process. The second compares alternatives and determines data, responsibilities and safeguards. The third develops a limited pilot with acceptance criteria. The fourth evaluates results and decides whether to expand, modify or withdraw.
The sequence is not purely linear. A review may require reformulating the purpose or choosing a different solution. Preserving that possibility signals good governance rather than project failure.
10.3 Responsibilities and evidence
Every significant decision must have a responsible person or function. Documentation should enable reconstruction of what data were used, which criteria were applied and which reviews were conducted. The aim is to support accountability and learning.
Evidence must include positive results and problems. An evaluation presenting only selected improvements prevents proper assessment of the solution. Known limitations, uncertainty and excluded uses must be communicated as clearly as benefits.
10.4 Periodic review
Periodic review examines whether the system remains suitable and whether its conditions of use have changed. Errors, complaints, accessibility, employment effects and costs must be reviewed. Frequency depends on risk and the pace of change in the process.
Incidents and user observations may prompt extraordinary reviews. The team must be able to stop specific functions without losing the entire service. The ability to intervene is part of design quality.
10.5 Professional judgment
The main contribution of Human Resources is to sustain responsible decisions about work and people. Technology can provide information, but it does not itself determine which goals are legitimate or how opportunities and responsibilities should be distributed.
An organization prepared for artificial intelligence can question, verify, explain and correct. Keeping people at the center requires translating that intention into verifiable procedures. Professional judgment is also demonstrated when a decision is made not to use a tool.
Questionnaires
Chapter knowledge & Human Resources dilemmas
Dilemmas receive no moral or employability score and are not used to assess job candidates.