BBC Sounds, music, radio, podcasts.

A busy Chicago hospital.

A weary doctor.

And before him, a patient breathing heavily in ragged gasps.

The doctor has been on shift for 14 hours.

He's struggling to think straight.

The patient shows ominous symptoms of a heart attack.

And yet, this could well be a simple case of panic.

The doctor must make a choice.

Does he send the patient home with a simple pill to calm them down,

or summon a cardiac expert diverting a medic away from other patients who also need their care.

Either will have potentially difficult, if not fatal, consequences.

After many of my shifts in the emergency department, I would just go home and lie in bed

just agonizing about all of these decisions that I made.

That is the voice of Dr. Zayad Obermeier.

He now works in machine learning at the University of California in Berkeley.

But as you can probably tell, once upon a time, he was a doctor himself.

You know, you work a shift one day and the next day a colleague comes up to you and says,

hey, do you remember that patient that you saw yesterday and your stomach just sinks

because you're the one who's responsible for a lot of really important decisions.

I think you're all too aware of how often those decisions go wrong.

You'll know by now that the advocates of artificial intelligence promise that there is an easier way.

...