AI blamed for failing to detect patient's critical condition, leading to death

A grieving family in Brazil is pointing fingers at a new artificial intelligence system, alleging it played a role in the death of a 32-year-old psychologist. Rebeca Molina was hospitalized for suspected gallstones and urgently needed an intensive care unit (ICU) bed, but her family claims the AI-driven platform Core-MG, used by the Minas Gerais state government, failed to classify her condition as a top priority.

Molina's condition reportedly worsened daily while she waited for an ICU bed. Despite doctors deeming intensive care essential, the digital platform allegedly assigned her a severity score of 6.8, which the family insists did not reflect the reality of her critical state. This perceived misjudgment by the algorithm led to a delay in her transfer to another hospital, approximately 300 kilometers away, where an ICU bed eventually became available.

Tragically, Molina died just hours after arriving at the second hospital. Her death certificate lists septic shock as the cause, though doctors continue to investigate the rapid decline. Her twin sister, Samela Cardoso Tenente Furtado, expressed profound distress, stating, "My sister is not a number. She is not just data from an AI system." Furtado voiced concerns that medical decisions, once firmly in the hands of physicians, are increasingly being ceded to statistical calculations performed by algorithms.

The Core-MG system was launched just last month by the Minas Gerais government, intended to expedite hospital bed distribution and triage patients based on medical urgency. However, this incident has ignited a fierce debate in Brazil about the ethical implications of relying on AI for critical healthcare decisions. The state government has denied that the AI system caused Molina's death, asserting that all processes remain under medical supervision and patient transfers are determined by clinical conditions and bed availability.