00:00:01
our guest today is
00:00:03
noam lemastricht lattar dean of the sami
00:00:07
ofer school of communication
00:00:09
at the interdisciplinary center in
00:00:11
herzeglia israel
00:00:13
we are going to speak to him about what
00:00:14
the robot journalists
00:00:16
will replace human ones noam thank you
00:00:19
so much for speaking with us today it's
00:00:20
a pleasure to be here again with you
00:00:22
marco
00:00:23
what a provocative topic you have
00:00:26
written about
00:00:28
in your paper what what exactly is robot
00:00:31
journalism well you know that the
00:00:34
computers have
00:00:35
helped journalists to write story to
00:00:38
write to find facts
00:00:40
since the since the middle of last
00:00:42
century there was
00:00:43
what we call data mining and analytics
00:00:46
data analytics
00:00:47
which helped journalists to find find
00:00:49
the
00:00:50
facts and do investigative journalism
00:00:54
so this is not new what is now
00:00:57
developing is that there are new
00:01:00
programs
00:01:01
artificial intelligence programs who
00:01:03
actually get the facts
00:01:04
and within fraction of a second write
00:01:07
the story
00:01:08
instead of the human journalists so this
00:01:10
is the second
00:01:12
part of the puzzle which really today uh
00:01:15
there are stories
00:01:16
uh written in forbes in in other
00:01:20
news magazines that are untouched by
00:01:22
human
00:01:23
human journalists we have in a very
00:01:25
freak very quickly
00:01:27
once there is an event or facts and
00:01:28
effects are being found
00:01:30
the program the artificial program
00:01:33
writes the story
00:01:34
and the the name of the journalist is
00:01:36
really the name of a robot there's a
00:01:37
company
00:01:38
called narrative science in illinois who
00:01:41
actually
00:01:42
already doing it and collected quite a
00:01:45
lot of money
00:01:46
for their as investors so is this a new
00:01:49
phenomenon or
00:01:50
has this been around for a while didn't
00:01:52
your research at mit in the 1970s
00:01:55
actually predict some of this
00:01:57
no i in my research predicted the use of
00:02:00
touch screens
00:02:02
uh which were later used by steve jobs
00:02:04
and i did the first
00:02:06
really studies on interactive television
00:02:09
where we studied how
00:02:11
providing people in the television
00:02:14
studio providing them with
00:02:16
devices similar to what today you can
00:02:17
use the touch phones
00:02:19
and they provide feedback and i studied
00:02:21
how allowing
00:02:23
publics to to feedback during to provide
00:02:26
feedback during the discussion
00:02:28
how it affected the group dynamics and
00:02:30
the discussion dynamics
00:02:32
we did not predict at the time robot
00:02:35
journalism at
00:02:36
all actually data mining is quite
00:02:39
quite uh it has developed in the past
00:02:42
20 years so how pervasive is robot
00:02:45
journalism
00:02:46
it's still it's still in the initial
00:02:48
stages uh
00:02:50
because the the programs are really
00:02:52
starting in
00:02:53
2010 but they are they're doing
00:02:56
very quick penetration and because the
00:02:59
robot journalist has some real
00:03:02
advantages
00:03:03
first of all he never forgets facts he
00:03:05
can do research very quickly
00:03:07
he never asked for a day off and he can
00:03:10
write a story with
00:03:11
within seconds so that the event that
00:03:14
the facts have been
00:03:17
they don't discovered deadlines and if
00:03:18
you write the program
00:03:20
correctly he doesn't he's not even
00:03:22
biased as you know
00:03:23
most journalists are biased about about
00:03:28
the stories
00:03:29
but the robot journalist if he
00:03:30
programmed correctly can be completely
00:03:32
unbiased
00:03:33
so he never forgets anything for example
00:03:35
there was an earthquake in california
00:03:37
and a robot journalist wrote the story
00:03:41
when the name of the journalist that
00:03:43
appeared in the story was already asleep
00:03:45
los angeles times by the way is using
00:03:47
this kind of robotic
00:03:48
storage writers for quite a while
00:03:52
how fast is it growing and what are the
00:03:54
factors driving the growth
00:03:56
this is a good question the the echo the
00:04:00
ecosystem which really
00:04:01
promotes this uh this uh robotic
00:04:04
journalism
00:04:06
is of course the uh the new the
00:04:08
multimedia
00:04:09
uh type of the development of multimedia
00:04:13
multi-platforms
00:04:15
the the slow disappearance of the print
00:04:18
it you
00:04:18
really basically a question of costs you
00:04:21
know that
00:04:21
advertising money today shifts towards
00:04:24
from the print media to the to the
00:04:26
internet
00:04:27
and because the the robotic journalists
00:04:30
are very very
00:04:31
cheap very efficient very quick and
00:04:34
because of the economics considerations
00:04:36
i i trust that the robotic journalism
00:04:39
will grow
00:04:39
very fast you know there's a next next
00:04:43
month there's a big convention uh a
00:04:45
journalist convention in the city of
00:04:47
philadelphia in israel
00:04:48
about 2000 journalists come down to a
00:04:51
lot
00:04:51
and i wrote the story that it's a future
00:04:54
story
00:04:55
for this convention about robot
00:04:56
journalism and the editor of the booklet
00:04:59
that it is this printing for this
00:05:01
convention wrote me
00:05:02
you know norm this is great because now
00:05:04
the journalists will have to start
00:05:06
working
00:05:07
when they will see the threat of this
00:05:09
robotic journalist that they can really
00:05:11
analyze facts very quickly and write the
00:05:14
story very quickly it will force them
00:05:16
into becoming innovative to do more
00:05:19
in-depth thinking more in-depth research
00:05:21
so the the optimists view this entry of
00:05:24
robotic journalism
00:05:26
as a new era where where really good
00:05:28
zonialist
00:05:29
will not disappear but they will be
00:05:31
forced to really
00:05:32
think again how they can be innovative
00:05:34
how can they be more in-depth
00:05:36
analysis so it depends how you see the
00:05:39
cuff how full of
00:05:41
empty uh i think it will improve
00:05:44
the threat of robotic journalism it will
00:05:47
be there i i guarantee
00:05:49
that in any newspaper electronic
00:05:51
newspaper or print newspaper that you
00:05:53
will purchase in the coming years
00:05:55
a major part of the stories will be
00:05:57
written
00:05:58
by robotic journalists but still people
00:06:01
will seek
00:06:02
human journalists because they always
00:06:04
have hopefully you will have this added
00:06:06
added value added innovation added
00:06:09
perspective
00:06:10
so i don't i see it as a positive force
00:06:13
in improving future journalism so since
00:06:17
you mentioned costs
00:06:18
i wonder what sort of a sort of business
00:06:21
model
00:06:22
uh can be uh put around this kind of
00:06:25
journalism
00:06:26
well you know that i wrote in the paper
00:06:29
and
00:06:30
it's a good question because i foresee
00:06:32
the new the new leaders of the newsrooms
00:06:35
will not be the experienced journalists
00:06:37
with the pipe
00:06:38
but will be the the the computer
00:06:41
engineers
00:06:42
and just the other day on october 17th
00:06:45
salzburg
00:06:46
salzburger i hope i say it correctly the
00:06:49
the publisher of new york times was
00:06:50
asked with insight
00:06:52
what what do you have done in light of
00:06:54
this digital
00:06:55
development in media and he said i would
00:06:58
hire more engineers
00:07:00
so i do see uh tremendous saving in
00:07:03
costs
00:07:04
in future newsrooms that are going to be
00:07:07
fully automatic uh but and really the
00:07:10
leaders
00:07:10
will be those journalists geeks who will
00:07:14
understand in media and will know how to
00:07:16
use data mining
00:07:18
and will know how to add value to what
00:07:21
the robot journalists can do but it's
00:07:23
going to be very efficient in another
00:07:25
another aspect of it today we are going
00:07:27
into in targeted
00:07:29
advertising so the robot journalist
00:07:32
will be able not only to write a story
00:07:34
but it will be able to
00:07:36
immediately send it to mochal if he
00:07:37
knows that mogul is interested
00:07:39
in certain type of information so we we
00:07:42
have here a complete
00:07:44
automation of news gathering news
00:07:46
analysis
00:07:48
news writing story writing
00:07:51
and targeting of of the information
00:07:54
uh that's fascinating to what extent do
00:07:58
social media platforms like facebook or
00:08:01
twitter
00:08:02
or linkedin factor into this phenomenon
00:08:05
well the social media social networks
00:08:09
uh help spread the word uh
00:08:12
of these all these new developments
00:08:14
today you know with especially with
00:08:16
twitter
00:08:16
uh once there are good efficient robotic
00:08:19
journalists in operation
00:08:21
the these social networks will spread
00:08:23
the word
00:08:24
very very quickly so it will expedite
00:08:26
the development of these processes
00:08:29
with a via facebook i think
00:08:33
i'm a more pessimistic about facebook
00:08:35
because i think
00:08:37
it will probably if it will not change
00:08:41
uh the way it is today it will probably
00:08:43
disappear
00:08:44
in fair in favor of the twittering the
00:08:46
twitter
00:08:47
kind of systems which also developed
00:08:49
into the video
00:08:50
uh today i i i think i'm
00:08:54
i'm on facebook i haven't looked at my
00:08:56
facebook
00:08:57
page i think for the past two weeks
00:08:59
because it's always loaded with
00:09:00
information loaded with dogs and friends
00:09:03
and
00:09:03
everybody is helping traveling all over
00:09:05
europe and all over america and i'm the
00:09:07
only one who's working
00:09:08
and so so i think the the present
00:09:11
the present structure of facebook
00:09:15
will have to change but twitter is in
00:09:17
the right direction
00:09:18
quick to the point so
00:09:22
social networks will spread the word of
00:09:24
of robotic journalism very quickly
00:09:27
how much of robot journalism
00:09:30
do you see in israel today the israeli
00:09:32
media
00:09:33
not much because the writing of the
00:09:38
the software the to write the stories
00:09:42
and to do the data mining is quite an
00:09:43
expensive israeli more
00:09:46
is in the direction of developing
00:09:48
startups
00:09:49
which uh point the direction to this new
00:09:53
new software new new artificial
00:09:56
intelligence
00:09:57
techniques but then this the startups
00:10:00
are being sold to
00:10:01
american companies or european so israel
00:10:04
is
00:10:04
in more in developing the the seed
00:10:08
of these innovations but to make it into
00:10:11
a large scale
00:10:13
so far israel has not been very
00:10:14
successful
00:10:16
in taking small ideas and really making
00:10:19
them large scale
00:10:21
so the media the media in the media and
00:10:24
israel
00:10:24
is a very very sad situation it's
00:10:27
controlled by government controlled by
00:10:29
politics
00:10:30
by by by
00:10:34
rich people by capital and the israeli
00:10:37
media today
00:10:38
is uh nothing to to write home about
00:10:42
it's quite sad yeah i agree uh well i'm
00:10:45
going to ask
00:10:47
two or three questions wearing the hat
00:10:49
of a human journalist
00:10:51
okay so
00:10:54
i'll try i'll try to answer as a human
00:10:56
as a human professor
00:11:00
so one of the things that i think
00:11:03
a lot of successful journalists share is
00:11:06
i think they have three qualities
00:11:08
the first is they have relentless
00:11:11
curiosity
00:11:13
the second is they have an innate
00:11:16
skepticism
00:11:17
especially about pr and the third is
00:11:21
that they have the ability to tell
00:11:23
stories simply and clearly in a way that
00:11:26
it can be understood by
00:11:28
a large number of people do you think
00:11:30
these are functions that can be
00:11:31
automated
00:11:34
sorry to say but yes all right
00:11:37
so a curiosity what
00:11:42
the first one was innate skepticism
00:11:45
curiosity skepticism and storytelling
00:11:48
curiosity uh
00:11:52
data mining and and software which
00:11:57
does data mining on huge data silos
00:12:00
uh is really the ultimate expression of
00:12:03
curiosity
00:12:04
you know in my article i tell about this
00:12:07
young ladies
00:12:08
from israel who is declared by the mit
00:12:12
technology review as a one of the future
00:12:14
scientists who discovered that if in the
00:12:16
rural area
00:12:17
in a poor area if you have one year you
00:12:19
have drought to be followed the
00:12:21
following year
00:12:22
with a flood then the third year
00:12:24
probably will be an
00:12:26
outbreak of cholera okay a journalist
00:12:28
with a high curiosity
00:12:30
and she accomplished that by looking
00:12:32
over huge amount of data
00:12:35
a human journalist cannot cannot
00:12:38
have this ability to scan over huge
00:12:41
amounts of data
00:12:43
but it was a human researcher who made
00:12:45
those connections she wrote the program
00:12:47
now this program can go to other data
00:12:49
sealers and discover
00:12:50
other things i think i said before
00:12:54
i don't see the the best human
00:12:57
journalists disappear
00:12:59
but the robots journalists will force
00:13:02
people
00:13:03
uh like you mochal
00:13:06
to be on the toes and to be more curious
00:13:09
and to do more
00:13:10
skepticism is one value which
00:13:13
uh it can be translated
00:13:16
to statistics and even a good program
00:13:20
can be skeptic about the results and
00:13:22
study it
00:13:24
but i see i don't i go back what i said
00:13:27
before
00:13:27
i don't see good journalists disappear
00:13:30
but i see
00:13:31
good journalists doing better job in in
00:13:33
line of what you have described
00:13:35
as skepticism curiosity i see it
00:13:38
i see fewer journalists but better ones
00:13:42
in light of robotic journalism and this
00:13:45
is what
00:13:45
wrote me the guy who who wrote the
00:13:48
who's the publisher of the booklet for
00:13:50
the conference he said
00:13:52
norm i'm glad you wrote this article it
00:13:54
will force journalists
00:13:56
to start becoming alive again you vocal
00:14:01
i'm sorry to say represent a minority
00:14:04
of journalists who write today
00:14:06
journalists today most journalists and
00:14:08
i'm saying it is the dean of school of
00:14:10
communication
00:14:11
most journalists are quite lazy you have
00:14:13
to provide them
00:14:14
with a story you have to buy them with
00:14:15
information and they they publish it
00:14:18
under their name
00:14:19
and they don't even go out to do
00:14:21
investigative journalism investigative
00:14:23
journalism is risky
00:14:24
it's costly it's dangerous and fewer and
00:14:27
fewer people are doing it
00:14:29
robotic journalings where robotic
00:14:32
journalism
00:14:32
i hope will bring this to life again
00:14:36
right the the other one quality and this
00:14:39
is more for editors and for reporters
00:14:43
uh is the is that editors
00:14:46
exercise editorial judgment
00:14:50
uh and i wonder how robot journalists
00:14:54
fit into that category also well you
00:14:57
know google news
00:14:58
google news were the first in i think
00:15:01
2002
00:15:02
uh have developed this uh google news
00:15:06
that crawled today over 27 000
00:15:09
news sources and they say they have pro
00:15:12
they wrote the program a how to exercise
00:15:16
judgment
00:15:17
they take an event for example and they
00:15:19
can study what is the
00:15:20
the probability of this event or caring
00:15:23
they
00:15:23
they bring they take the criteria that
00:15:27
good editors are exercising and they put
00:15:30
it into the programs
00:15:32
it's like a good chess player the a good
00:15:35
chess program will never study
00:15:37
and will never improve itself by by by
00:15:40
gaming against a weak a weak partner so
00:15:44
what the people who write these programs
00:15:46
go to the best editors
00:15:48
they they follow their best practices
00:15:51
and interview them and they they tried
00:15:54
to write the program
00:15:55
to ask the same questions that a good
00:15:57
editor would ask
00:16:00
so in conclusion let's end with the
00:16:03
question that we asked at the beginning
00:16:06
can robot journalists replace human
00:16:08
journalists
00:16:10
what the answer i'm getting from you is
00:16:12
yes but not completely
00:16:14
yes but not completely but there is one
00:16:17
pessimistic angle to this
00:16:19
robotic journalism from point of view of
00:16:22
economics is very efficient
00:16:24
and a very good journalist would be
00:16:26
quite quite expensive
00:16:28
and i fear that economic considerations
00:16:31
among in the news media
00:16:34
board of directors could force more and
00:16:37
more
00:16:37
the new media organization to go into
00:16:39
robotics because of the cost advantages
00:16:42
uh as a human being i i hope
00:16:46
that at least a certain portion of every
00:16:48
media in the future
00:16:50
will be run by by high quality
00:16:52
journalists
00:16:53
that will will make sure that the human
00:16:56
journalist is always better
00:16:58
than the robot journalist but it is a
00:17:00
hope
00:17:02
well maybe someday uh both human
00:17:05
journalists and human professors
00:17:08
uh will become better and robot
00:17:10
journalists and robot professors
00:17:12
will do the work that humans don't want
00:17:14
to do norm thank you so much for joining
00:17:16
us today thank you it's a pleasure
00:17:17
michael thank you very much and i hope
00:17:18
you
00:17:19
continue with your good journalistic
00:17:20
work thank you the human journalistic
00:17:23
work
00:17:23
thank you
00:17:47
you
