Jyunmi Hatcher

All right.

There was a team, a junior research team that consisted of a master's student and a high school kid empowered with AI tools.

They used AI to build medical prediction models,

in just a few minutes, that match or beat what teams of PhD-level data scientists spent months building.

And the results of that got published in a major medical journal.

So...

This paper came out February 17th in Cell Reports Medicine, and it's been making the rounds.

So, all right.

So why does this matter?

So first, some context.

This had to do with preemie or preterm birth, babies born before 37 weeks.

It's the number one cause of newborn death in the United States.

About a thousand babies are born too early every single day.

Globally, it's 11% of all babies born or preterm each year.

And researchers have been trying to research why this happens.

It's not locked down on the causes.

So researchers at UC San Francisco built one of the largest pregnancy data sets in the world.

They pulled together microbiome data, that's the bacteria in the body, from about 1,200 pregnant women across nine different studies.

So a huge data set.

So the problem with having such a huge data set was analyzing all the data.