A supercomputer at Beth Israel Deaconess Medical Center in Boston, Massachusetts, is analyzing patient data to predict medical outcomes, including the likelihood of death. According to Dr. Steve Horng, project lead at Beth Israel, the system can forecast a patient's probability of dying within 30 days with 96 percent confidence. "If the computer says you are going to die, you probably will die within the next 30 days," Horng said.
The system pulls from doctors' visits, lab results, medications and vital signs, then compares current cases against data from more than 250,000 individuals treated over the past 30 years. That expanding database allows the machine to apply machine learning and produce a rapid diagnostic assessment, flagging present and future risks such as heart attacks, infections and cancers.
How the prediction works
Patients are connected to the machine, which processes their available medical history in real time. The supercomputer does not rely on a single test; it weighs patterns across decades of prior cases to estimate what is likely to happen next. The result is a probability, not a certainty, though Horng described the confidence level for mortality within 30 days as near 96 percent.
Ethical concerns and the project's stated purpose
The concept of a "death predicting machine" has renewed questions about the ethical and societal consequences of artificial intelligence—concerns previously raised by figures including Elon Musk and Stephen Hawking. Horng countered that the project's aim is to use big data to support doctors, not supplant them. "Our goal is not to replace the clinician; this artificial intelligence is really about the doctor's ability to take care of patients," he said.
The source material does not specify how the 96 percent figure was validated, whether the system is in active clinical use, or what patient consent and data privacy measures are in place. Those details remain unaddressed in the available information.
Editorial Notes