AI systems used in triage of patient symptoms, or any health application for that matter, that are based the creation of individual profiles - profiling techniques - even if temporary in nature, fall under data processing mechanisms for high-sensitive data. They are forbidden in principle, and possible under two relevant exceptions:
1. The provision of valid consent to be subject to automatic decision – this means Explainable Artificial Intelligence (XAI) methodology provisions need to be clear, and doubts about which healthcare settings/and patients characteristic create conditions for an open, comprehensive and free consent exist. The right to object is an additional issue. This exceptional route does not seem to be the most viable option in healthcare, as trade-offs between consenting and receive care, especially in emergency situations, may raise additional legal and ethical issues.
2. If this data processing is necessary for reasons of public interest, according to EU or national law, in proportion to the aim/purpose, respect the essence of the right to personal data protection and if the adequate and specific measures are ensured to safeguard the fundamental rights and the interests of the data subject, which may include technical developments within XAI methodologies still somehow underdeveloped.
1. What sort of XAI methodologies would you believe could ensure open, comprehensive and free consent to be obtained?
2. Are there clinical circumstances, and or groups of patients to whom, such online services are restricted by nature of the unclear balance between free consent and the clinical conditions, i.e. patient having a severe migraine or stroke, is he/she in the condition to fully appreciate and reason, the elements of the consent he/she is giving?
Comentários