Skip to content
Home » Ethical Considerations in AI Mortality Risk Prediction: Insights from INTERVENE

Ethical Considerations in AI Mortality Risk Prediction: Insights from INTERVENE

    Artificial intelligence (AI) is increasingly being used to predict health outcomes, offering new possibilities for early intervention and personalised care. A recent study from Finland introduces an AI model capable of estimating the risk of death within one year for every resident in the country, a development that highlights both the power and the ethical complexity of predictive healthcare.

    The research analysed data from more than 5.4 million people, incorporating over 8,000 health, demographic, and socioeconomic factors. The model proved highly accurate, outperforming traditional approaches based on age or sex. By capturing signals of biological frailty, a state of vulnerability beyond chronological age, it represents a new kind of “digital ageing marker.”

    The analysis showed that while the model achieved strong overall accuracy, its performance varied across demographic groups, with predictions being most reliable for younger women and less accurate for older men and disadvantaged populations, including low-income individuals, migrants, and people with mental health conditions.

    While the study showcases the potential of AI to identify health risks at scale, it also underscores the need for caution and responsibility in its use. The authors stress the importance of inclusive data collection, transparent model design, and ongoing evaluation to ensure that AI systems benefit all social and demographic groups equally.

    As a proud INTERVENE partner, CPE advocates for responsible, transparent, and patient co-designed AI in healthcare, ensuring innovation enhances access, trust, and equity while reflecting patients’ real-world experiences.

    Learn more about the project here.

    INTERVENE_Article on AI