Hi everybody Cheryl Ackison here.

Cheryl Ackison here.

Welcome to a special from the Archives edition of Full Measure After Hours.

Today, can artificial intelligence step in and fix America's bad polling problem?

You do not have to dig too far back, most of you at least, to think about the 2016 election and remember how far off the polling was.

Donald Trump was polling last in a lot of the early polls

and even toward the end it was widely said by just about everybody that Hillary

Clinton was a shoe-in. Hillary Clinton was a shoe in.

Hillary Clinton was literally said to be measuring the drapes in the Oval Office

prior to the election because it was such a sure thing.

Afterwards, when Donald Trump won despite pretty much all the media's predictions that he didn't have a chance.

And by the way, pretty early and often, I think I was one of the few or only in

national media that was predicting based on anecdotal evidence I was gathering

across the country based on watching the rallies of Trump, Hillary Clinton, and

Bernie Sanders. I was one of the fewer only ones predicting from the start that he

was going to win. When Donald Trump won there was a lot of effort by the media to try to revise history a little bit and say well we never really predicted that and if you look at the numbers we kind of said the numbers were there all along and the fact is and I did some stories on this for full measure

The media and many pundits and polls were wildly off

So there was a lot of discussion you may remember post 2016 about how we can rely on polls and I did some reporting on the fact that many of you know this.

Polls have now become a tool of propagandists rather than just measuring public opinion.

They seek to influence public opinion.

To make you think a candidate might be unpopular when he or she isn't, or to make you think that a particular

candidate is more popular than he or she is.

...