Ethical Aspects of Work Disability Risk Prediction Using Machine Learning

Our study focuses on the areas of social and economic sustainability in machine learning. The risk of work disability can be predicted with machine learning and using various data sources. Machine learning techniques appear to be a potential tool to support expert work and decision-making. We will present the five stakeholders of the work disability prediction—the employee, the employer, the occupational health care, the pension fund, and society. All these stakeholders should be taken into account when developing AI to support disability risk prediction. We will compare two methods with different data sources, occupational health care data, and pension decision register data. There is still another stakeholder, the data scientist, who is developing the machine learning algorithms. We will present five important aspects of the data processing and algorithm design phase: non-maleficence, accountability and responsibility, transparency and explainability, justice and fairness, and respect for various human rights. These aspects need to be considered when collecting data, storing it in databases, and sharing it with others.