Would you let algorithm help you decide if a loved one in coma should die?
Except that's not what the article states at all.
"Kaufmann agrees — a computer system’s assessment should only matter if it determines that a coma patient shows promise after human doctors deemed it a lost cause — not the other way round."
In other words, if doctors say there is no chance of recovery, and the computer says there is, treatment will continue. If doctors say there is a chance of recovery, and the computer says there isn't, treatment will also continue. It's only being used as a tool to confirm decision making when both doctors and machine agree there is no chance of recovery.
Also, this is largely hypothetical. It's based on functional MRI. Many centres don't offer functional MRI, so those patients are out. Even in the centres that do, there are a huge number of reasons you can't MRI patient because of metalwork, implants, etc, so those patients are out too. And even in the patients that can have an MRI, transferring an ICU patient to the MRI is difficult and risky - if doctors were going to continue treatment anyway, there would be no benefit to exposing the patient to the risk.