Triage 4.0: On Death Algorithms and Technological Selection. Is Today's Data- Driven Medical System Still Compatible...

Area of Science:

• Health Informatics
• Artificial Intelligence
• Machine Learning

Background:

• Health data offers significant promise for improved health outcomes and efficiency.
• Machine Learning (ML) and Artificial Intelligence (AI) leverage big data for advancements in healthcare.
• However, the use of health data also introduces vulnerabilities and potential risks.

Purpose of the Study:

• To explore the dual nature of health data utilization in ML/AI.
• To identify potential negative implications and ethical concerns associated with big health data.
• To emphasize the need for caution and robust criteria in AI/ML healthcare applications.

Main Methods:

• Review of potential benefits and risks of big health data in ML/AI.
• Analysis of ethical considerations including discrimination, eugenics, and social Darwinism.
• Discussion of the long-term implications and lack of established criteria for safe implementation.

Main Results:

• ML/AI in healthcare can support diagnosis, treatment, and cost-effectiveness.
• Undesirable side effects include discrimination, 'mechanisation of death', and selection with eugenic potential.
• Current methods are insufficient to prevent severe negative outcomes from big health data.

Conclusions:

• Handing over decision-making to machines in healthcare is potentially dangerous and irresponsible.
• The use of AI/ML with health data conflicts with human rights and constitutional principles.
• Further development of criteria and long-term experience is needed to mitigate risks.