WEBVTT

00:00.680 --> 00:03.240
We have a very fragmented media ecosystem.

00:03.240 --> 00:05.240
We have multiple social media platforms.

00:05.240 --> 00:08.160
We have news coming from many different places.

00:08.160 --> 00:09.720
One of the biggest concerns for this

00:09.720 --> 00:13.400
is the ease by which misinformation and disinformation

00:13.400 --> 00:16.440
can spread very quickly from one platform to another,

00:16.440 --> 00:18.200
making it quite difficult to detect.

00:23.200 --> 00:24.840
Tools like artificial intelligence

00:24.840 --> 00:26.720
can be used for both good and bad purposes.

00:26.720 --> 00:28.800
And that's something that I think researchers

00:29.440 --> 00:32.040
in society as a whole are only realizing now.

00:37.680 --> 00:39.560
Fact checkers, I think, are really necessary

00:39.560 --> 00:42.200
to help identify good and bad information,

00:42.200 --> 00:45.080
but computers are really fast at being able to detect it.

00:50.600 --> 00:52.320
What are the tools we can build

00:52.320 --> 00:55.360
that can help fact checkers do their jobs more easily,

00:55.360 --> 00:58.200
that allow them to handle the overflow of misinformation

00:58.200 --> 00:59.120
that comes at them?

00:59.120 --> 01:01.800
How can we use natural language processing

01:01.800 --> 01:04.240
to make them to be more effective in their work,

01:04.240 --> 01:08.240
rather than trying to automate the important decisions?

01:13.160 --> 01:15.040
We'll never get to a place

01:15.040 --> 01:18.360
where our media environment will be absolutely perfect.

01:18.360 --> 01:21.880
The success of good systems are, yes, better systems,

01:21.880 --> 01:24.640
but ultimately we need to have better people

01:24.640 --> 01:25.920
and better society.

01:25.920 --> 01:27.960
Whether you're a journalist who wants to build

01:27.960 --> 01:29.440
trust within your newsroom,

01:29.440 --> 01:31.840
or you're a researcher who's trying to encourage

01:31.840 --> 01:33.240
trust in science,

01:33.240 --> 01:35.360
I think there's a lot of people who care about this problem

01:35.360 --> 01:37.560
and wanna work together to find solutions.