5. Automation of Judgment and AI: Ethics and the Future of Work

1. The article analyzes the evolution of artificial intelligence from the automation of simple tasks to the automation of judgment—namely, processes of classification and selection. The author emphasizes that AI models provide only probabilistic signals, while the final decision and the determination of the acceptable error threshold remain a normative and ethical choice. Using examples of errors in medical algorithms, it is shown that data are not neutral but reflect existing social inequalities. The text argues that algorithmic fairness is not a technical problem to be solved with code, but a matter of consciously defining values and success criteria, which requires maintaining human oversight (human-in-the-loop) and critical reflection on delegating decision-making competencies to machines.