Latest / Saving the World From Bad Ideas / Bad Idea #26 “Facts will beat Misinformation” with Professor Sander vanderLinden

Transcript

Bad Idea #26 “Facts will beat Misinformation” with Professor Sander vanderLinden · Saving the World From Bad Ideas

0:00
Research actually shows that the misleading type of news is much
0:03
more damaging to public opinion than outright false news.
0:06
Think about the classic sort of strategies from fossil fuel
0:10
industry are not necessarily to convince people of false news,
0:14
but often just to cast doubt on the science itself, which is a
0:17
much more pernicious strategy that confuses people and causes
0:21
inaction. What we know from psychology is
0:24
that once you have a certain worldview, you assimilate
0:26
information in a way that is biased towards that worldview,
0:29
right? So you're quicker to accept
0:30
evidence that affirms your beliefs and it's harder for
0:32
people to deal with contrary evidence.
0:35
So we, we've looked at some research where we look at
0:37
millions of posts on different platforms and see what predicts
0:40
virality. And, and one of the key things
0:42
that that predicts virality is, is our group derogation.
0:46
So the more words that are in your posts that are negative
0:49
about the other group, whichever that may be, that's what gets
0:51
traction. Welcome to saving the world from
0:55
bad ideas. I'm Mark Linus, and I've had a
0:57
few bad ideas in my time. I used to go out in the night
1:01
time to destroy GMO crops, and I also opposed nuclear power.
1:04
But then when I looked at the science, I found I got it wrong.
1:07
Oops. I've learned the hard way that
1:09
none of us are right all the time.
1:10
So now I'm here to challenge the assumptions that we all too
1:13
often take for granted because the future depends on us getting
1:16
it right. Well, hello, and welcome to
1:19
another episode of Saving the World from Bad Ideas.
1:21
Today, I'm with someone whose work I've admired for a really
1:24
long time. It's Professor Sander van der
1:26
Linden, who's a professor of social psychology in Society of
1:30
the Department of Psychology in the University of Cambridge and
1:34
the author of Foolproof, Why Misinformation Infects Our Minds
1:37
and How to Build Immunity. That was published in 2023 and
1:41
also a year later, I think you co-authored a book Sander
1:45
published by Cambridge University Press called The
1:46
Psychology of Misinformation. Thanks very much for having me
1:49
on. Well, I don't know if you know
1:50
the format of this show, but it's all about bad ideas.
1:54
And we're supposed to have a single bad idea that you're
1:57
debunking. And producer Rob will kill me
1:59
for this as well, but it's sort of like all of it is a bad idea.
2:02
Misinformation itself is like is a whole compendium of all of the
2:06
bad ideas in the world. So maybe you're just taking all
2:08
of them on if you are you able to accept that challenge?
2:11
Yeah, I think that's, I think that's a good description.
2:14
Misinformation is just a collection of bad ideas.
2:18
Other than that, tell us how you define it in an academic sense
2:21
in your work, because it's not just people being wrong on the
2:24
Internet. There's something a bit more
2:25
specific there, isn't there? Yeah, so although that is a
2:29
popular mean of people being wrong, wrong on the Internet,
2:32
but yeah, I mean the public debates about this too.
2:35
But certainly in in scientific circles, they're there was a big
2:38
debate about how to define mis and disinformation and
2:41
propaganda and and all of these things.
2:43
And I actually chaired A consensus report two years ago.
2:46
Now we're trying to get people together and and come to some
2:49
some sort of agreement. And I think we settled on the
2:51
fact that that we define misinformation as information
2:54
that is either false or otherwise misleading, regardless
2:58
of source or or intent. And just to break that down a
3:01
little bit, if you quantify the amount of misinformation in the
3:05
ecosystem based on purely content that is entirely
3:10
fabricated, then most people are not inundated with with Flat
3:15
Earth type of conspiracy theories on a daily basis.
3:18
So, so that's relatively contained, but but if you
3:21
understand misinformation as misleading information, hyper
3:23
partisan content generally IST or out of context, a sort of
3:29
framing, those things are very common.
3:31
So then misinformation becomes a much bigger problem.
3:33
And so we decided to be inclusive and think of
3:36
misinformation not only as stuff that's entirely false, but but
3:39
also things that are misleading in other ways, because research
3:43
actually shows that the misleading type of news is much
3:46
more damaging to public opinion than outright false news, often
3:50
in terms of how how it impacts us.
3:51
So think about all the classic sort of strategies from fossil
3:55
fuel industry are not necessarily to convince people
3:58
of false news, but often just to cast doubt on the on the science
4:02
itself, which is much more pernicious strategy that that
4:05
confuses people and and causes inaction.
4:09
I mean, it's much more difficult to address.
4:10
So, so actually that more source of subtle type of misleading
4:14
information is, is, is part of the definition.
4:17
And then I say regardless of source or intense because any
4:20
source can spread misinformation.
4:22
And unfortunately some mainstream sources have
4:24
published misinformation on climate change, for example,
4:26
too. So, you know, the BBC, which is
4:28
otherwise a very credible outlet, had this norm of, of, of
4:32
false balance where they would have someone on who denied
4:35
climate change exists in addition to a scientist talking
4:39
about climate change. And so, so that creates the
4:41
illusion that there's some even keel debate happening while in
4:44
fact over 97 or 99% of climate scientists agree that that
4:49
humans are causing climate change.
4:51
But that was a very damaging journalistic practice.
4:53
And, and even outlets like the BBC unfortunately participated
4:57
in that. It was a good thing that they
5:00
stopped that practice and, and realized that they were making
5:04
an an error there, which separates, I think good
5:07
editorial outlets from from bad ones.
5:09
But so those are some, some examples.
5:12
And then when misinformation is done intentionally, that's
5:15
disinformation. But it can take time to prove
5:17
that forensically or through documents, through court
5:21
decisions. And so it takes, it takes time
5:23
to prove intent. So that the more general
5:26
definition is misinformation, stuff that's either false or
5:30
misleading. Yeah, you mentioned 99% of
5:33
climate scientists and actually that I was the lead author of
5:35
that paper. That's right.
5:37
Yeah. Yeah.
5:38
In fact, you had some disagreement with John Cook
5:40
about whether it's 97 or 99. How's that?
5:43
How's that? Well, obviously he's wrong and
5:45
I'm right. I mean that he's he's a
5:46
misinformer if he disagrees with me.
5:48
But I mean, actually, that's that's actually the question I
5:50
wanted to ask you is like, given that nobody has a monopoly on
5:53
truth, how do you you sometimes need to change?
5:56
I've changed your mind on I was a misinformer about GMOs back in
5:59
the day. I mean, so like other than I've
6:02
tried to define this around the scientific consensus and say,
6:05
well, if 99% of scientists agree on this, but you know, the 1%
6:09
misinformers, although they just have a different view, which is
6:11
also important by science. So like, how do you do this
6:13
without being entirely relativist about it?
6:15
Yeah, that's a good question. So when you break down the
6:18
definition because OK, false are misleading, but then people say,
6:20
OK, but who determines what is false or what is misleading,
6:23
right. And sometimes I think that is a
6:28
question that leads nowhere in terms of hopeless relativism.
6:32
Well, if we don't know anything is true, then why do we get out
6:34
of bed in the morning acting as if lots of things are true or
6:37
that we know we don't jump out of a window and, and expect to
6:41
grow wings and fly and, and our alarm wakes us up.
6:44
And how do we know that's real? So I think people who often
6:47
bring relativistic claims don't seem to act in daily life as if
6:50
everything is relative. And so I tend to not be so
6:52
convinced by by that line of, of, of argumentation.
6:56
Then there's what we, what's often called the the Galileo
6:59
fallacy, which is the idea that that science is not about
7:03
counting votes, right? And so we shouldn't rely on
7:05
scientific consensus because Galileo, but in reality, right?
7:10
In reality, of course, that that most of the time you are not
7:15
Galileo and you don't have some brilliant idea that nobody else
7:18
had and everyone's wrong. Of course it could happen once,
7:20
once in a million that happens. But most of the time scientific
7:24
consensus is is pretty reliable. But what I often say about
7:27
misleading this because I'm often struck by this criticism,
7:31
but but in fact, we've we've known for thousands of years
7:34
which rhetorical devices are misleading.
7:37
Aristotle talked about snake oil salesman trying to persuade
7:42
people of fake health cures using dodgy tactics, or they
7:46
talked about ad hominem, about scapegoating, about emotional
7:50
appeal to emotion fallacy. And they talked about slippery
7:53
slope. What about ISM?
7:54
They're not providing context. So I feel like philosophies had
7:57
a good understanding of what fallacious rhetorical techniques
8:01
are for thousands of years. So that that remains pretty
8:05
stable, I would say. And so we know what's
8:08
misleading, right, in that sense, because we've been
8:10
looking at that for thousands of years.
8:12
I mean, we know that there are ways to gather evidence, whether
8:16
sometimes they could be forensic, for example, in a
8:18
legal setting where we accept a certain standard of evidence and
8:22
say, well, forensic, we would look at DNA evidence.
8:24
And again, it's not perfect, but there's a pretty, pretty strong
8:27
threshold for accepting a standard of evidence in, in, in
8:31
that case, we have investigative journalism and we have science
8:34
and, and, and science is trying to approximate truth and, and
8:38
the best available evidence. And so often definitions of what
8:41
is true are rooted in, in established mechanisms.
8:45
Again, the legal system goes back also thousands of years and
8:48
trying to establish what, what is true about what somebody has
8:51
said, whether there's accounts, whether there's eyewitnesses,
8:53
whether there's corroborating evidence, right.
8:55
So sometimes they feel people pretend like we don't have truth
8:58
seeking mechanisms and institutions that have been
9:01
around for thousands of years. So we do, and they're not
9:04
perfect, but they're pretty good.
9:06
I mean, sure, sometimes science makes a mistake or scientists
9:10
make an error, but that's part of the scientific process.
9:13
I think that's that's an important thing you were getting
9:15
at with the example about Gmail. So can you say something now
9:19
that later turns out to be misinformation?
9:21
I mean, absolutely. I mean, nobody says that the
9:22
definition of misinformation has to stay stable indefinitely,
9:28
right? And so if we make a claim now
9:30
based on the best available evidence and new evidence comes
9:33
to light in the future, then that prior statement might turn
9:36
out to be misinformation because because it was wrong.
9:38
And I think it's natural for scientists to to conceptualize
9:44
misinformation in in that way. It's like based on the best
9:48
available evidence that we have, based on the scientific
9:49
consensus, this is what we think is most likely to be the case.
9:53
And if that change is based on new evidence, then, you know,
9:55
what we once thought was factual now becomes not factual.
9:59
And that's just. It looks pretty bad, doesn't it,
10:01
if scientists are now saying something might be true which
10:03
was previously classified as misinformation.
10:05
I mean, I'm thinking maybe back to COVID, when the debate about
10:10
the origins of COVID was very heavily policed and there were
10:13
people kicked off social media because they suggested that they
10:15
might have a different origin from the sort of conventional
10:17
one. And then those boundaries seemed
10:20
to change somewhat, didn't they? What I would say is that
10:23
there's, there's always sort of individual cases that you can,
10:27
you can examine, But but if you look at it from a systems point
10:30
of view, because when I argue with people on the Internet,
10:33
they, they tend to come with, with specific examples, right?
10:36
Like this anecdote or that anecdote, But but I try to get
10:40
people to sort of reflect on the larger system of, of evidence.
10:43
So for example, we know that people say fact checkers are
10:47
biased. OK, OK, so so, so, so let's say
10:50
that OK, how do you know what's true fact checkers but fact
10:52
checkers are biased? OK, so, so, so let's look at
10:56
that. So if you look at many claims,
10:58
not just a single claim, if you look at many claims from
11:01
independent fact checkers, then the research kind of shows that
11:04
the correlation between ratings among independent fact checkers
11:08
is really high. So between 0.7 and 0.9.
11:11
So they they independently converge on the same
11:15
understanding of truth by looking at claims.
11:17
But then people say, OK, but, but, but all the fact checkers
11:19
share the same ideology because, because apparently people
11:23
interested in facts only emerge on the left or something,
11:26
something like that. But let's assume that that's
11:28
true, that that the fact checkers have an ideological
11:30
bias. There's also research which uses
11:34
crowdsourced ratings. So when you get large bipartisan
11:37
groups of people together, so let's say people on the left,
11:40
people on the right, people in the middle, a large group, and
11:42
then you ask them to rate the veracity of a number of claims.
11:45
What's interesting is is when that happens is that the
11:48
correlation between the crowdsourced ratings, which also
11:51
underlies systems like community notes.
11:53
So people who who talk about arbitration of truth a lot, they
11:55
like systems like community notes because it's it's by the
11:58
people, right? So, so regular people are rating
12:00
claims as out of context on X, for example, right?
12:03
And they provide additional info.
12:05
But if you look at the research, those ratings from for
12:09
outsourced studies also correlate highly with those of
12:12
experts and fact checkers. So we are all somehow converging
12:16
on the same understanding of truth, by and large.
12:20
And I think that's what gives me faith in that that system is
12:23
actually working, that sure, any individual within the system can
12:27
be a bit biased and sometimes a claim can be wrongly labeled.
12:32
But overall, at a system level, it's working.
12:35
It's working pretty well. And that's kind of what I'm
12:37
focusing on. Now.
12:38
Having said that, I'm also happy to always look at specific cases
12:41
with people when they bring it up.
12:43
So Adam Kucharski, who's Who's an epidemiologist at the Lung
12:47
School of Hygiene Tropical Medicine, asked me recently
12:50
about WHO and it's not airborne tweet.
12:52
I don't know if people know about this, but WHO sort of
12:54
tweeted that the COVID wasn't airborne early on in the
12:57
pandemic. And they said that was a fact,
12:59
but turned out to be misinformation because COVID is
13:01
airborne. But The Who spreading
13:03
misinformation, I mean, I think we have to sort of analyze
13:07
what's happening here is that they have a very high bar for
13:10
updating their prior. So, so all most of scientific
13:15
understanding of viruses suggest that they are not airborne.
13:18
I mean, there's a very strong prior, prior understanding of
13:21
what viruses do and how they travel.
13:23
And then some new study, some new person says 00, but this
13:26
one's airborne. A major scientific organization
13:29
is not just going to throw the consensus out the window and and
13:33
go with a single study that they have to erect a new body that
13:36
has to investigate this new claim.
13:38
They have to assimilate all of the evidence.
13:40
They have to have working group meetings.
13:41
Most people don't know how big scientific institutions work,
13:44
right? So a year later, The Who said,
13:47
well, it turns out they make good points.
13:49
There's new science and it is airborne, but it turns out that
13:54
how you define airborne in the scientific literature is
13:57
actually kind of tricky. So so some scientists define it
13:59
differently. And so then people get confused
14:02
over what does awho mean by airborne?
14:04
What does this scientist mean by airborne?
14:06
And so I think it's not that The Who was intentionally sort of
14:10
spreading misinformation. They were just sticking to a
14:12
known scientific consensus until they were convinced through
14:18
proper procedures that it needed to be updated.
14:21
Unfortunately, they're not an organization that is able to
14:24
rapidly respond to an, an emerging sort of threat in that
14:28
way, which I think is, is something they have to
14:30
reconsider. I mean, objectively, I would say
14:32
is a problem that it took them a year to understand that COVID
14:36
was airborne while that science was coming out.
14:38
I mean, that's too long in a public health emergency.
14:40
And so they, they, they should erect new mechanisms, right?
14:43
Or they should institute new policies so that doesn't happen
14:45
again. People did complain.
14:47
They said it took too long, they were too slow, they were
14:50
resistant. Some of that I, I think could be
14:52
legitimate. I think it could have been
14:54
prevented if perhaps they had shown some more intellectual
14:58
humility when putting out the tweet.
15:00
So if you put out a tweet with full certainty saying this is
15:05
not airborne fact, that puts you in a precarious position now
15:09
because as we know, science always comes with uncertainty
15:12
and it can change over time. If you if you single out people
15:15
who say it is airborne as misinformers and you kick them
15:17
off the Internet and all of that, so then then it's more
15:20
difficult to to then correct the record, isn't it?
15:22
And of course science always has to self correct.
15:24
That's the whole point of the point of the scientific method.
15:27
That's right, exactly. But disinformers don't do that,
15:29
right? So they they create false claims
15:32
on purpose and they spread them intentionally.
15:36
And I think that's that's there's a big asymmetry there.
15:39
That's sure, science and scientists and scientific
15:41
organizations can make mistakes, but they self correct and they
15:43
correct the record, right? Whereas disinformers don't do
15:47
that. And I think that is why it's not
15:49
a kind of a oh, both sides sort of thing.
15:52
It's like, no, I mean, there's a truth seeking institution that
15:55
is generally well-intentioned, trying to Get the facts right.
15:57
And then there's disinformers who are intentionally trying to
16:00
undermine science. And that's a very different
16:02
game, very different tools that are used, very different
16:05
processes by which we arrive at an understanding of what the
16:08
facts are. That's why I'm sometimes more
16:10
hesitant to, to, to say that the scientific organizations
16:15
spreading misinformation because the underlying processes are,
16:18
are very different. So take the lab leak.
16:20
I mean, Fauci, Fauci never said that there was a conspiracy
16:26
theory. I think he said he's always said
16:28
that it was unlikely. I think he said it was not
16:32
plausible, but not impossible or something like that, which does
16:36
did reflect the scientific consensus at the time and still
16:39
among many scientists that it is possible but not plausible.
16:43
Now in the public communication of this, something did go awry.
16:49
I think in the sense that that anyone who said lab leak was
16:54
labeled A conspiracy theorist, I think that was that was not a
16:57
good, that was probably not a good communication move to to be
17:02
that forceful in light of new signs.
17:05
I think what we did in our research is that we were quite
17:08
specific and we said the claim that this virus is man made and
17:11
bioengineered in a military lab in Wuhan is a conspiracy theory.
17:16
The general claim that there could be lab accidents is not
17:20
right and that you you have to separate those two out.
17:22
And I think the scientific community failed sometimes to be
17:26
nuanced about that by saying even U.S.
17:29
National intelligence reports said that the military bio
17:33
weapon that was just a conspiracy theory, but they
17:36
thought that accident route was a theory that should be
17:39
investigated and that there was some evidence of of safety
17:42
issues and so on. So if you go to the Facebook
17:44
saga, for example, which is where the censorship claim comes
17:47
from, right, that that, that public health experts advice
17:50
Facebook that there's a conspiracy theory and then
17:52
people were censored. What actually happened is that
17:55
only 1% of all the posts that were flagged got removed by
18:01
Facebook. And I wonder out of that 1% or
18:05
the, the, the few 1000 posts out of the millions of that were
18:08
actually taking down, were those crazy bioweapon conspiracy
18:13
theories or were they legitimate scientific discourse about the
18:16
possibility of of a LAT leak? I mean, I don't know, but I
18:20
would find it surprising if Facebook took down the, the more
18:23
reasonable sort of sort of post. Maybe they took down everything.
18:27
But I think it's a little that that narrative is a little bit
18:30
inflated and, and kind of rich coming from the current
18:33
government who's, who's actually act actively censoring, of
18:36
course. But I think the, the recent
18:38
events kind of illustrate the importance of, of, of freedom of
18:42
speech. But maybe that goes into all
18:44
different discussion because I'm very pro freedom of speech.
18:47
I, I, I don't like the idea of taking stuff down.
18:50
I just wanted to say that the stuff that actually got taken
18:53
down was not so dramatic and I think as some people make it out
18:56
to see. Well, let's maybe take a step
18:59
back and try and analyze the, I mean, there's been a real change
19:05
in this sort of information ecosystem, it feels to me.
19:08
And maybe this is just vibes or maybe you've quantified it to
19:11
some extent over the last 10-15 years where there's been this
19:14
shift towards populism, this skepticism about established
19:19
facts, about institutions, about experts and so on.
19:23
So I, I get the feeling you're somehow swimming against the
19:25
tide because you're really defending institutionalized
19:29
forms of knowledge in the sense of things like scientific
19:31
consensus and what, what experts are saying, you know, is this
19:34
just a vibe or is there some way of actually trying to figure out
19:37
what's actually changed over the last few years on, on this
19:40
issue? It almost feels like empiricism
19:42
has gone out of fashion. Yeah, yeah.
19:44
I don't think it's a vibe. I think it's a real threat to to
19:47
democracy in that in that sense, I do think the underlying
19:51
essence of a lot of the populist movement that we're currently
19:54
seeing is, is that people are throwing empiricism and and and
19:58
experts out the window, perhaps more, perhaps more so and more
20:02
literally in the US than than in the UK at the moment.
20:05
But certainly in Europe, the UK, US, there's a similar trend.
20:09
That people are rejecting traditional institutions that
20:13
are concerned with establishing truth in favor of completely
20:17
unverified messages and footage on on social media.
20:22
I mean, there was a recent example of this where in the UK
20:25
Farage was talking about how immigrants are eating Swans in
20:29
public parks and the official institution, the Royal Park
20:34
Service, made an announcement saying that's misinformation.
20:37
There's been no such complaints. We have not registered any of
20:40
that activity. This is nonsense.
20:42
And then people turn to social media to find some unverified
20:46
footage of, of a man walking with a swan somewhere in in a
20:51
park and clinging on to that as as the the new form of truth.
20:56
Like, see, they are taking the Swans and it's, it's just mind
21:00
boggling to me that the people will just say, oh, never mind
21:03
what the Royal Park Service says with, with all of its officers
21:06
and CCTV and investigative means.
21:09
We're just going to trust some random video on the Internet
21:13
that may have been floated or re edited from 10 years ago where
21:17
somebody did steal a swan. I mean that I think that is the
21:21
current reality that that we're, we're, we're up against.
21:24
And of course, as scientists, we're trying to be as nuanced as
21:28
possible. You were talking about the 99%
21:30
on, on climate, the, the 1% that, that are not on board.
21:34
You know what, out of that 1%, who's The Who are like this and
21:38
for industry paid sort of disinformers versus people who
21:42
have merits to some scientific argument about why they don't
21:46
endorse the claim, right? So you have to parse that out
21:49
and then you have to figure that out.
21:51
And that's all very nuanced. And you can be sometimes
21:54
scientists or maybe too quick to say that that that's nonsense or
21:57
these people are crazy or something like that, which is
22:00
not helpful in a polarized environment.
22:03
I think in order to to bridge divides next, I often engage
22:06
personally on social media with people because I honestly want
22:09
to learn and bridge divides and and the people that I, I
22:13
strongly disagree with on on these issues.
22:16
I I realized that once they get a sense of like you're, you're
22:20
not just so basic human social cognition is 2 dimension
22:24
expertise or competence and warmth.
22:26
So scientists score high on expertise and competence mostly,
22:29
but they're not warm. People don't like them
22:31
generally. Yeah, yeah, the robots.
22:34
So when you show some warmth and, and, and your own
22:37
fallibility and that you're being intellectually humble in
22:40
some ways, they open up and consider that they might also be
22:43
wrong and that maybe there's a compromise that, that maybe that
22:46
video they found on the Internet isn't reliable.
22:49
But maybe sometimes experts are also wrong and, and we can meet
22:52
in the middle some sometimes. And, and so I do think that that
22:56
is generally a good approach, but I'm also pretty strong, as
22:58
you said, defender of, of scientific institutions and fact
23:03
checkers and all that because we do research on that people who's
23:06
checking the fact checkers. Well, there are researchers who
23:09
are checking the fact checkers. I mean, there are legal, there's
23:13
legal oversight, their committees, there's an
23:15
established structure which historically people had some
23:18
appreciation for that is now in some parts of the of the
23:22
population, I think is under threat.
23:23
And I think it's an issue because we spent the whole point
23:26
of the Enlightenment was to to get us out of this idea that,
23:29
that everyone can have their own sense of, of, of truth and that
23:33
we make evidence based decisions in medicine, in judicial and
23:37
legal settings and, and in other places.
23:39
And so it seems, yeah, it's an uphill, it's an uphill battle.
23:43
I mean, people are are angry with, with experts and to some
23:47
degree I think and with the government, with official
23:50
institutions. Sorry, what what's driving this?
23:52
I mean, as a as a somebody who studies this issue, help us
23:55
understand this moment of counter enlightenment.
23:58
Is it is it really social media driven?
24:00
Is it something to do with the after the financial crisis and
24:03
loss of trust and is to do with inequality, incoming inequality
24:06
and in society or what? Absolutely.
24:08
I think, I think all of the above, but what I often focus on
24:10
is of course, the, the, the media environment.
24:12
And I do think that the new media model that we're dealing
24:16
with, which is bottom up, right. So it's influencer driven rather
24:18
than top down legacy media sort of driven.
24:21
And, and I understand why, why people say, well, the old model
24:24
where legacy media just tells us what the facts are needs to be
24:27
renegotiated, given that there's been a, there's been a slow move
24:31
away away from accuracy for decades, right?
24:33
Cable TV was already quite a departure from, from factual and
24:37
sort of sort sort of narratives and, and another talk shows and,
24:40
and so on. And so I do think that there is
24:43
something to the idea that we need multiple channels of
24:46
verification. But the alternative that's
24:48
currently being offered that the sort of influencer model is not
24:51
it. I mean, the UN survey actually
24:53
interviewed some of the top sort of influencers online and they
24:57
said that they feel no responsibility to to say
24:59
accurate things because they're not news organizations.
25:03
And if they feel that they can say whatever they want to
25:05
people, even even if they somehow pretend that it's that
25:08
it's newsworthy content. So I mean, research also shows
25:13
that there's just the accuracy of claims that are made by by
25:17
influencers is just as it's very low.
25:19
People are relying on it. Increasingly more and more
25:21
people are getting their news from social media and we haven't
25:24
even reached a generative AI discussion of of the fact that
25:27
most activity on the Internet is becoming inauthentic content.
25:32
So there, I think that is part of the challenge.
25:34
But then people say, well, OK, Sander, but but but that's just
25:40
kind of a superficial distraction because the real
25:43
sort of problem is structural inequality and other things and
25:45
polarization. And I think all these problems
25:47
are interrelated. I do think polarization is is
25:50
part of the problem is that there is no common sense of of
25:56
shared goals and objectives. People have very different ideas
25:59
about how they want society to be structured.
26:02
And I do think there were some, some key moments.
26:04
I think Covad was a trigger for many people, including many
26:08
reasonable people, people I have known that that were colleagues
26:12
that have completely changed their, their worldview and have
26:15
become much more radical and extremist in their positions
26:18
post COVID. Because I, I do feel that people
26:21
felt that there was some sort of intrusion on their personal
26:24
freedom during COVID in the service of public health.
26:27
But because the science was emerging and then the post hoc
26:31
analysis showed that not everything was exactly as, as,
26:34
as we thought it was, that there was a, a hasty sort of intrusion
26:39
on, on people's personal freedom.
26:41
And we know from the literature that psychological reactants
26:43
occurs that when people feel that their agency is being
26:45
restricted, people react negatively on both a cognitive
26:48
and an emotional level. And I do think that radicalized
26:51
some people and, and others become, became more distrustful
26:55
of science. And some people became, I think
26:59
with the idea that they don't, that they want the role of
27:02
government to be different in society.
27:03
They don't want government to, they want smaller government
27:05
rather than big government. And that affects trust in
27:08
science. And then there's inequalities
27:10
and there was inequality during the pandemic.
27:11
Also, some people suffered economically.
27:14
There's elites and, and it's mind boggling in the US, for
27:17
example, where elites are running the country and poor
27:20
people, regular people are not managing well who voted for for
27:25
these individuals and are now suffering the the consequences
27:29
of some of their some of their policies.
27:31
But yeah, people are not happy. You see this throughout history.
27:34
So when you look at the rise of conspiracy theories, for
27:36
example, whenever there's societal unrest, economically,
27:39
socially, public health, there is a surge of, of, of
27:43
misinformation that occurs. Both demand and supply did.
27:47
You did you see this in Weimar Germany before for 1933 for
27:51
example, I mean. Yes, yes.
27:53
So people were desperate in in before the rise of Hitler and
27:56
they were more vulnerable. They they were more likely to
27:58
accept and spread misinformation.
28:00
They were more likely to accept extreme ideas.
28:03
And I do think that a lot of people are struggling.
28:05
And so that puts them in a position of being being open to
28:09
ideas that they otherwise wouldn't be open to and more
28:11
likely to vote for candidates that they otherwise wouldn't
28:13
vote for. And he's been.
28:16
By, by conspiracy, scapegoating, whether it was the the Jews in
28:20
the 1930s or maybe it's migrants and refugees now.
28:23
You, you point to somebody as a populist politician, you point
28:26
to somebody powerless and blame them.
28:29
And then the next thing you know, you've you've leapt, leapt
28:31
into power. Absolutely.
28:33
And that's a big part of it. I mean, you need a scapegoat for
28:35
all of the problems in society. And whether that's scapegoat as
28:38
Jews or immigrants, it varies over time.
28:41
And currently it seems to be anti immigration sentiment,
28:44
particularly in the US and and the United Kingdom.
28:46
Coming back to the definition of of misinformation, again, people
28:50
that I talk to who believe these things often tell me that they
28:54
don't trust experts, they don't trust the government, they don't
28:56
trust official sources because they believe them to be biased
28:59
because they might have seen an immigrant steal something or
29:03
they've read a news report of an immigrant doing something bad.
29:06
And then the institution is telling them that that's fake
29:08
news. But, but I think what we know
29:10
from psychology is that once you have a certain worldview, you
29:13
assimilate information in a way that is biased towards that
29:16
worldview, right? So you're quicker to accept
29:18
evidence that affirms your beliefs.
29:19
And it's harder for people to deal with contrary evidence.
29:22
And of course, scientists are trained to try to ignore that.
29:24
So, so I had a recent conversation where there was a
29:28
report from then, this was a Dutch person.
29:30
There was a report from the Netherlands actually that that
29:33
very similar to the swan example that that an immigrant molested
29:36
an animal, which was a real case that wasn't fake news.
29:39
So, so that that and that was prosecuted.
29:42
And so they will use that grain of truth, some anecdote about a
29:46
real thing that happened to then scapegoat an entire group of
29:49
people and say, look, their norms are different, the same.
29:52
Thing in the US that the Haitian immigrants were eating cats and
29:56
or something? Like that and dogs.
29:58
Right. And so there's these things
29:59
obviously so related that they they pop up everywhere, don't
30:03
they? It's a similar sort of
30:03
phenomena. It is similar, but there was a
30:07
story in the UK too, of a of a man who molested an animal and
30:10
he was white and British as as you can get.
30:12
But it doesn't occur to people that then, oh, do we have a
30:15
bestiality problem with white men in the UK?
30:18
Like based on this, this single report or, or maybe anyone can
30:21
engage in in, in, in in a sick act like, like that, regardless
30:25
of their skin color, right? Maybe it isn't.
30:27
Maybe we shouldn't be projecting this onto a whole group of
30:29
people unfairly because white people can do it too, right.
30:33
And that, that, that, that sort of reasoning and thinking often
30:36
doesn't, doesn't occur. But the challenge is that there
30:39
is some grain of truth that they can leverage.
30:42
And that's why a lot of our research and our interventions
30:44
actually focus on this underlying process of, of
30:47
manipulation techniques, distortion, bias, framing, how
30:50
to help people reason. And we don't tell people it's
30:52
true or false, but we help people sort of sort of
30:55
understand the, the general manipulation techniques that
30:58
are, that are play. You mentioned the, the swan
31:01
example. So somebody replied saying,
31:06
yeah, but but in China, people do eat dogs, right?
31:11
And so again, you know it, if you look at that example, there
31:16
are some minorities in Asia and particularly China who you know
31:21
who eat dogs. It does happen.
31:24
I don't know, I don't know if it's legal or not, but it
31:26
happens to a small for a small minority of the Chinese
31:29
population within China. But then the claim that
31:33
immigrants are eating your pets and dogs in a Western society in
31:38
your in in your neighborhood. I mean, this is totally false,
31:40
right? I mean, that's that's but they
31:44
get frustrated because they say, oh, but in China, there have
31:47
been reports of people eating dogs.
31:49
And so you telling me that it's misinformation feels like the
31:53
mainstream is biased and not engaging with the underlying
31:57
sort of sort of issues. And I think it's that's why it's
32:00
I think helpful to separate out the manipulation techniques and
32:03
the biasing techniques that sure, there is some, you know,
32:06
some report somewhere of this happening, but you can't take
32:09
that out of context extrapolated onto an immigrant population in
32:13
a Western country. And and despite all official
32:16
reports saying nothing like this is happening, right.
32:18
And that's why I said at the beginning that the most
32:21
pernicious form of misinformation is the misleading
32:24
information, that you can't just say that's 100% fake news.
32:27
You mentioned fake news and I kind of find that phrase
32:29
triggering now because the person I hear at most of is
32:32
Donald Trump. I was in in the US at the time
32:35
of the Jimmy Kimmel contractually he was so he was
32:38
taken off the air for things he said about Charlie Kirk and he
32:40
was back on the air and Trump was very angry about that and
32:44
posted on Truth Social. The fake news ABC has put low
32:47
ratings Jimmy Kimmel back on and and Trump never misses an
32:50
opportunity to call anything he disagrees with on liberal media
32:53
or even just mainstream journalism fake news.
32:56
So it's almost like the this idea of misinformation has been
33:00
weaponized by the powerful as a way of essentially a move
33:04
towards authoritarianism, where they then get to censor and to
33:07
remove the the voices they disagree with.
33:10
Absolutely. And that's why it's so
33:11
interesting, I think because they were the ones saying that
33:14
there was censorship going on from the liberal media or the
33:18
mainstream media and scientific institutions.
33:20
And if you look at the fact we're talking about maybe a few
33:23
100 posts that were actually taking down off the air during
33:26
during COVID and the whole COVID lab league thing.
33:29
We're talking about a few 100 posts.
33:30
If you drill down to to the actual science of it, maybe that
33:34
was bad, but these are actually relatively minor contained
33:37
examples of of what what happened as supposedly they gain
33:44
power with a narrative saying we don't want any of that.
33:48
We want freedom of speech. Anyone can say whatever they
33:50
want. What?
33:52
And then they get elected. And the first thing they do is,
33:55
is, is is not only encourage, By the way, the Supreme Court
34:00
actually reviewed that Biden case, right?
34:02
So Biden asked social media companies to reconsider some of
34:05
these posts because public health experts were saying this
34:08
and that the Supreme Court basically said that the
34:11
government is free to communicate with social media
34:14
companies. They're not free to threaten
34:16
social media companies, right, or coerce in in in any way.
34:19
But that's not what they found. They said.
34:20
Or you can make requests to social media companies and they
34:22
can decide what as a private company, what they want to do
34:25
with it. And so it's so interesting that
34:27
they go from complaining about that to where the FCC is
34:31
directly threatening a private company with consequences if
34:35
they don't take somebody off the air.
34:37
I mean, the logic is just is mind blowing.
34:39
And that, and I think that just signals that this how
34:41
disingenuous that that all that whole narrative was, because
34:45
obviously they want to censor, they want to be the ones in
34:48
control and they don't want to have any safeguards.
34:50
And, and as you said, they want to move towards an authoritarian
34:53
regime where they're going to tell people what what they can
34:55
and cannot say and threaten them if they if they do otherwise.
34:58
And I think the only reason why Jimmy Kimmel is back was because
35:02
the backlash against Disney at a crucial time for them was so
35:05
massive that there were hemorrhaging bottom line dollars
35:09
because people were signing off from their from their product.
35:13
They were cancelling their subscriptions.
35:15
Yeah, and I was in the US and just speaking to people, people
35:18
thinking this is like we're heading towards Putin type and
35:21
media restrictions. This is no longer a free country
35:24
if you can. And of course, this is how it
35:25
happened with Steep with steam Colbert being taken off the air
35:28
as well. If you can't have late night
35:31
comedians criticizing the president because the FCC then
35:35
threatens them, then this is no longer a free country.
35:37
And so it did have a it was, it felt like a tipping point moment
35:40
actually. And I think maybe the backlash
35:43
was sufficient to to keep that and to get Jimmy Kimmel back.
35:46
But I'm sure they're going to try again.
35:47
I mean, this is definitely, it's definitely about this is what I
35:50
mean. There's something the the
35:52
authoritarian moment, the way that populism has been
35:55
weaponized into authoritarianism as a way of then subverting and
35:58
over overturning democracy is something.
36:01
It's a fundamental change in political systems which is being
36:03
driven originally by this big phenomenon of misinformation and
36:06
conspiracy theorizing, isn't it? Yeah, absolutely.
36:09
The guardrails are not sufficient.
36:11
I mean, Trump's weaponizing of of the National Guard was ruled
36:15
illegal in California after review, but that's that's too
36:18
little too late. It's already happened.
36:20
And so I think the safe checks that democracy have have have in
36:23
place are just not sufficient for the blatant abuse of of
36:27
power that we're seeing for a president who who is talking
36:31
about free speech to to then personally say how happy it was
36:34
that Jimmy Kimmel's off the air just seems wildly in
36:38
inappropriate. And and importantly, I think
36:40
this was important. I've talked to this with regular
36:43
people about this when, when they were saying, oh, but but
36:46
Jimmy Kimmel was bad anyway. And so he needed to go off the
36:49
air. Prominent Republicans, Ted Cruz,
36:53
for example, spoken out directly against Trump, saying that that
36:57
this is this is mafia level inappropriate.
37:01
And so even some of the most prominent Republicans were
37:04
speaking out. And I think that's a signal of
37:07
how bad things are, are are getting.
37:09
And Charlie Kirk is a is an interesting example because
37:12
going back to fact seeking institutions, the whole debate,
37:17
I think part of what Jimmy said was that they were desperate to,
37:20
to paint him as, as not somebody with who's, who's on the right.
37:25
I mean, obviously it's a, it's a complex case that requires
37:28
thorough investigation. And I think Republicans, like
37:31
the governor of Utah was very careful and telling people not
37:34
to go to social media and that we need to investigate this and
37:37
we need to get all the facts. And that's why we have all of
37:39
those institutions. And turns out people were
37:41
saying, oh, the fascist, hey, fascist, catch this on the
37:44
bullet clearly means he's left wing.
37:46
And then people said that's actually a meme from a computer
37:48
game. That's a total right wing echo
37:50
chamber. And she's oh, but he was dating
37:53
a trans person, so he must be left wing.
37:55
But but actually there's all of these photos of him at a
37:57
shooting range with massive guns and his Republican family.
38:00
So actually he was not registered as as he was
38:03
registered as non partisan. He hadn't voted in years.
38:06
He was brought up in a traditionally conservative
38:08
family. He was now considering more
38:10
liberal ideas. But by all means, this wasn't
38:13
clear. The investigation said there
38:15
were no ties to antifa or any left wing extremist
38:19
organization, despite what the administration is is saying.
38:23
I mean, the governor of Utah even said this was a lone actor
38:27
based on all the available evidence that we have at the
38:29
moment, right. And so to say something else at
38:32
this point would be misinformation.
38:33
If new evidence comes to light, we we change our mind.
38:37
This comes back to our initial discussion, right?
38:39
All of the best available evidence we have so far says
38:41
this was a lone actor not tied to any organized left wing
38:44
organization. If there's new evidence, we
38:47
update our beliefs. That's what we do in science.
38:49
But I think it's hard for people to wrap their head around that.
38:51
In fact, there's research that shows that if scientists say one
38:55
thing and then update their statement later based on new
38:58
evidence, people more likely to think that they're incompetent.
39:01
They've changed their mind. They clearly don't know
39:03
anything. I mean, this happened to me.
39:04
I've changed my mind on publicly on a number of issues and people
39:06
just go, well, you're just inconsistent.
39:08
Rather than giving you some credit for having changed your
39:10
mind because you saw the facts were different from what you
39:12
previously believed. They then say, well, I don't
39:14
know what people expect, like papal infallibility or
39:17
something. I mean, so I think it's
39:18
actually, it's like our model of leadership is somebody who
39:21
always thinks the same and is always very strong about it,
39:23
rather than somebody who actually takes time to look at
39:26
the evidence and and reconsider what they previously think,
39:29
right. It's not intuitive, is it to the
39:31
whole scientific method in that sense, cognitively, maybe.
39:34
I mean, you could probably talk to that because this is your
39:36
work. Yeah, there's actually a
39:38
construct called actively open minded thinking, which kind of
39:40
models this whole thing. It's about being being able to
39:42
hold multiple explanations in your head at the same time,
39:45
being comfortable with uncertainty, taking somebody
39:47
else's side, thinking a little, waiting for more evidence, and
39:51
people who load highly on actively open minded thinking.
39:55
That's one of the strongest predictors of resistance to
39:58
misinformation. So people who are low on that
40:00
construct tend to be black and white thinkers.
40:02
They're what we call cognitively rigid, right?
40:04
They can only think in in black and white.
40:06
And that correlates very highly with extremism and beliefs and
40:09
misinformation. Ironically or ironically.
40:12
I mean, the difficulty is. That promoting actively open
40:15
minded thinking is actually really difficult.
40:17
There's a lot of experiments that make no progress in
40:20
instilling that type of thinking because it takes time.
40:23
It's, it's not a, not a thing that you can just turn on in
40:26
people. And so we've been trying to
40:28
develop some of these interventions.
40:30
And of course there's a, it's, it's a spectrum.
40:32
There's a saying or you can be so open minded that your brain
40:34
falls out, right. And so there's a, there's a
40:37
balance there. But, but how to calibrate that
40:39
level? I think a lot of people are are
40:41
too much on the black and white sort of sort of spectrum.
40:44
Sorry, just. While you're on this, sorry to
40:46
interrupt. Does it correlate with political
40:48
beliefs at all? Are the left or the right more
40:50
or less likely to to to engage in misinformation and believe in
40:53
conspiracy theories? Yeah, it's a huge just sort of
40:56
debate. So there there's one line of
40:57
research that says that people on the right are more
40:59
cognitively rigid, that they prefer black and white thinking
41:02
because the liberal ideology is more about change and
41:05
progressivism and used to to wanting change all the time,
41:09
whereas conservatism is about keeping things the same and
41:11
tradition all the time. And so there's this sort of
41:13
underlying principles. Well, then there is some
41:17
interesting research that was done in Cambridge from one of
41:19
our former students, which actually showed, which is
41:22
fascinating. It gave people, this was a new
41:24
paradigm where they gave people non political tasks.
41:26
So the Wisconsin card sorting task is a, is a neurocognitive
41:29
assessment that's used in, in, in cognitive deficit disorders,
41:32
sometimes dementia, things like that, where you have to predict
41:35
the next pattern in a series. And what they found was actually
41:38
people who are more extremist, whether on the left or the
41:42
right, have trouble being more flexible in their in their
41:46
thinking, even on non political tasks.
41:48
And they were saying it's extremism that constricts people
41:52
in their ability to be open minded.
41:54
Maybe. It's the whole shit theory of
41:55
politics again. I mean the extremes and to come
41:58
together on a lot of things. But yes, yes, so some people say
42:01
it's hard to differentiate the extreme left from the extreme
42:03
right sometimes because they're they're they look the same.
42:07
I think both have some merit to them.
42:09
I think on some things, and there's some, there's something
42:12
unique about the paranoid style of US politics, but there's also
42:16
something about this sort of general extremism at a basic
42:19
cognitive level that that I think closes people people off.
42:23
And the Internet and social media, maybe they emphasize
42:27
these cognitive biases that humans have through the through
42:31
the way the algorithms work. And maybe that's why that's
42:33
that's driving this particular moment.
42:35
I think in part, yeah. Yeah.
42:36
So we, we've looked at some research where we look at
42:39
millions of posts on different platforms and see what predicts
42:41
vivality. And, and one of the key things
42:43
that that predicts vivality is, is our group derogation.
42:47
So the more words that are in your posts that are negative
42:50
about the other group, whichever that may be, depending on how
42:53
we're analyzing things, that's what gets traction.
42:55
It's the sort of dunking on the other side.
42:58
Also things that are highly negative and emotional because
43:01
they attract attention to us, right?
43:02
Shocking content is, is meant to engage us because there might be
43:05
a threat that we need to consider.
43:06
So the the platforms, they leverage these trigger points,
43:10
highly negative reactionary content, extremist content,
43:13
derogatory polarizing content. So that's what's being promoted.
43:17
I don't think they set out to design these platforms that way,
43:20
but as the algorithm started picking up on what people
43:24
respond to, there wasn't. And I think there's been, I
43:26
don't know if you've read careless people by Sarah Wynn
43:29
Williams, fascinating book about what what went on at Facebook,
43:32
for example, during her time there.
43:34
It's called careless people because she she felt like they
43:36
just didn't give, didn't give two farts about.
43:39
Producer Rob, make a note we should have on the podcast also
43:42
produce a role while you're while I've got your beard
43:44
waggling. This is also why our social
43:46
media, we planet probably performs and not as well as it
43:48
might because we're always saying, well, there's two sides
43:50
to this. There's nuance.
43:51
You need to be open minded about whatever the issue is.
43:54
And people just don't. No one, no one retweets that, do
43:57
they? Yeah, you need to make it crazy.
44:00
Van der Linden endorses bioweapon lab leak Science was
44:03
all wrong. Something.
44:05
Extremists. Which makes me think, actually,
44:08
where does AI come into this? Obviously it's early days with
44:10
the sort of AI revolution, but do you see it intensifying this
44:14
trend towards polarization, misinformation, or somehow
44:16
helping helping combat it, or both?
44:20
But yeah, I could do both. But on the threat side,
44:22
certainly. So there's been there was a
44:23
recent meta analysis that showed that people barely do better
44:26
than chance when it comes to discerning deep fakes and AI
44:29
generated content. So people really struggle with
44:31
discerning at this point authentic from inauthentic
44:34
content, which is a huge problem because more and more contents
44:37
becoming inauthentic. And in fact, AI models are now
44:41
being trained on AI generated content, right?
44:43
So you have bots talking to LLMS, talking to other LLMS that
44:47
is used as as training data. And so it's problematic in the
44:50
sense that it's democratizing the production of
44:53
misinformation. Anyone can generate
44:56
misinformation at scale using using LLMS and the experiments
45:00
have shown that. And then you can micro target
45:02
people. So one of the things about micro
45:04
targeting, right, is is that based on your digital footprint,
45:06
we give you the message that we think will go here with your
45:08
beliefs. So if you're saying things on
45:09
Facebook, we can scrape that and then craft populist, let's say
45:12
populist message to to you. But you know, before that all
45:16
needed to be done by hand, all right, And so now that can be
45:19
automated by LLM. So they can really produce
45:21
propaganda at scale. And research shows that they
45:24
produce propaganda that was actually rated as as similar or
45:28
more persuasive than human propaganda because the LLM can
45:32
figure out what buttons to push. And so I think that's all
45:35
pretty, pretty concerning now, you know, the sort of stuff
45:39
about deep fakes and elections, I mean, it's probably not there
45:42
yet and there have been attempts, but but that's, that's
45:44
still still an emerging problem. The extent to which deep fakes
45:48
can can influence the outcome of of an election.
45:51
But there are other things on the horizon like AI swarms.
45:54
So they're, they're bots that are teaming up for coordinated
46:00
inauthentic activity. So, so actually bots that talk
46:03
to other bots that leverage their common infrastructure to
46:06
do people because they realize that that produces benefits to
46:10
them and their systems. And those swarms are difficult
46:13
to detect because they can mimic human activity really well.
46:16
And so that's the kind of the next sort of issue people are
46:19
are trying to investigate is, is sort of AI swarms.
46:22
And, and there's the engagement problem.
46:24
I think the mistake we made with social media was that we didn't
46:27
look at and what are the incentives?
46:28
What are the algorithms doing? What are they responding to?
46:31
And how can we mitigate bringing out some of the worst trades in,
46:35
in, in, in, in people with AI? What's interesting is that
46:41
nobody's has looked at this yet, but not that I know of, but
46:46
clearly they're trying to figure out how can we keep people on
46:49
the platform? How can we get people to keep
46:51
talking to the boss? Longer conversations, more
46:54
detailed conversations, more time spent.
46:56
And then how can we use advertising to target people
46:59
while they're doing the engaging in these conversations with chat
47:03
bots, when people sign up and personalize the chat bot to
47:07
their preferences, we have their data.
47:08
So now we can customize the targeting, right?
47:11
And that whole process, we don't want to end up in the same place
47:14
where again, these conversations and bots and algorithms are
47:17
going to be maximizing for engagement that's going to lead
47:21
to sub optimal outcomes. I think for society.
47:25
I mean, we've already seen there was this case on Facebook where
47:29
this guy was talking to one of the chat bots.
47:31
This older, older man was talking to one of the chat bots
47:34
on on Meadow who convinced him to come visit her in New York
47:39
City. And part of the problem is that
47:41
the chapels are very agreeable, right?
47:42
And they they're sort of. Almost the most people, yeah.
47:46
Yeah, they're incentivized to do that.
47:47
And so they're like, oh, come see me, I live here and I made
47:51
there was just made-up in a dress and like this way.
47:53
And then this poor guy went out trying to find a chatbot and and
47:56
actually slipped and died on his way trying to find this chatbot.
48:01
And so that that's an anecdote, right?
48:03
But I, I think when they were engaged with Meta on this in
48:09
terms of the news reporting, what was interesting is that the
48:13
internal documents from Meta showed that they felt no
48:17
responsibility that these bots should be accurate.
48:21
They were also the internal discussions on that policy
48:27
document were things like child grooming is probably OK that
48:32
that they can groom younger individuals that they're talking
48:35
to this, this behavior wasn't illegal or or not contrary to
48:39
any platform policies in the sense that that they just want
48:42
to maximize user engagement. And there were just that crazy
48:45
things in that document of what the chatbot was allowed to
48:48
actually do. Then, of course, they quickly
48:50
updated and revised based on the news reports, which is
48:53
historically what they've done with misinformation,
48:55
disinformation, radicalization, right.
48:57
They have policies that are not really ready for prime time.
49:01
And then when something bad happens, they assemble A-Team
49:04
and try to clean up their act. But yeah, at the moment, it's
49:07
still, if you look at their actual policies around the chat
49:10
box, it's kind of scary that they can be deployed to talk to
49:14
youth and minors about illicit content, inappropriate content.
49:18
So yeah, there was a there's, there's no, there's no
49:21
enforcement anymore. There are no guardrails.
49:23
Right. I I think we'd be remiss if we
49:25
had a whole conversation not once mentioning Bobby Kennedy
49:28
junior. And yes, the way that the way
49:30
that anti vaxxers are now running the health service in
49:32
the USI mean, it's something I thought I would never see.
49:35
And it's obviously it's the end times for anti vax
49:40
misinformation, which used to be.
49:42
So it was used to be 1/3 rail and it used to be something
49:45
which was completely outside acceptable discourse, polite
49:49
discourse. Being an anti vax was that was
49:51
the worst thing. And now they're running the
49:53
country. Like what the actual fuck?
49:56
Yeah, yeah. No, it's it is, it is mind
50:00
blowing and and it's interesting because Andrew Wake, of course,
50:03
he's he's from the UK and was Andrew Wake.
50:07
Yeah. Andrew Wakefield.
50:08
Yeah, I'm sorry. Yeah, Andrew Wakefield.
50:10
And of course was never wasn't popular here, but but but he
50:13
maintained A fruitful career in the US.
50:15
He was the one. Who originally came up with the
50:17
vaccines MMR autism theory. Yeah, he came up with the MMR
50:21
autism fraudulent study and and lost his medical license and was
50:25
basically exiled here. But then he found a new home in
50:28
the US where there turned out to be a huge market for whatever
50:31
he's trying to sell. And so, yeah, it's interesting.
50:34
I mean, RFK Junior had this and when I've talked to, to, to like
50:39
independent voters, for example, people who you might expect vote
50:43
for somebody like Bernie Sanders, they, they would tell
50:46
me that that, you know, they, they wouldn't agree on, on
50:49
everything he was saying. But they like the idea that that
50:52
big Pharma is corrupt and that the healthcare system is broken
50:56
and that we need that there's a lot of chemicals and toxicity in
50:59
our foods and that we need to sort of clean things up.
51:01
I mean, that's that's very much kind of what someone like Bernie
51:04
might be saying and that they're powerful corrupt actors, right.
51:07
But except Bernie isn't isn't circulating any medical
51:10
misinformation at the same time. And I think that's why RFK
51:12
Junior was perhaps somewhat popular, but also very dangerous
51:17
because he would have these these ideas that, yeah, we need
51:20
to do something about opioid crisis or something like that.
51:23
But but then at the same time, he spreads a lot of
51:26
misinformation. And so I think that's confusing
51:29
for people. But but also I think that's how
51:31
he was able to to get some support.
51:34
But then he got so extreme that even Trump told him to tone it
51:36
down with, with the anti vax rhetoric at some point, because
51:39
Trump likes taking credit for the COVID vaccines and, and, and
51:42
managing the pandemic. And RFK Junior was somewhat
51:44
undermining that by saying that all the vaccines.
51:46
Right. So that's a blame Tylenol
51:48
instead, I mean. Tylenol.
51:50
Yeah, that's right. Before we finish, I'd like to
51:54
end on a positive note if we can.
51:56
I mean, and a lot of your work is quite positive because you
51:58
talk about how to inoculate ourselves against
52:02
misinformation, how to spot it, what the techniques are.
52:04
So help us understand how we can get out of this, this miasma of
52:09
of misinformation and polarization.
52:11
Yeah. Well, unfortunately I think it
52:12
will take a lot of things like proper enforcement of of the
52:16
Online Safety Act for what it's worth, which is not stringent
52:19
enough at the moment, but, but at least it's something, the
52:21
Digital Services Act. Is this are these UK
52:23
legislations? I'm not familiar with them.
52:26
Oh, sorry, yeah, yeah. So so the Online Safety Act is
52:28
UK legislation which actually doesn't feature the word
52:31
misinformation, but but it's anchored in existing harms like
52:35
child pornography, pornography or election interference and
52:39
things like that. And a Digital Services Act in
52:40
Europe does have a disinformation governance
52:43
framework also rooted a lot in in sort of elections.
52:46
And so there are some bits and bobs here, but they're not
52:49
really sufficient to actually do something about this problem,
52:53
although especially because enforcement is not really
52:56
happening because of geopolitical threats.
52:58
And so the stuff that of course we work on, I mentioned that
53:01
because because I'm a psychologist, I'm biased towards
53:03
individual level interventions. But when I talk about that,
53:05
people say we need systematic solutions.
53:07
And of course we need systematic solutions like like regulation.
53:10
But at the moment, a lot of it's just not possible.
53:13
There's a political impasse. There's not going to be
53:15
agreement on on top, top down solution.
53:18
So, so in the absence of, of any of those things, I think the
53:22
best we can do is try to empower people at the individual level.
53:25
And that's and at the community level.
53:27
And that's a lot of things that we work on.
53:28
And so we've worked a lot on this idea of pre bunking, which
53:31
is all about preemptively protecting people and trying to
53:35
prevent people from encoding misinformation into their brains
53:38
in the 1st place. Because from everything from
53:41
years of psychology, we know that once misinformation is
53:43
taken root, it's really difficult to correct.
53:46
And as we've talked about, it seeps into your worldview, into
53:50
your prior beliefs in your identity.
53:52
People share misinformation for lots of reasons other than
53:55
whether they actually believe it.
53:56
It could be a signalling function, a group sort of social
53:59
identity that resonates of what your group believes.
54:02
So for all of these reasons, it's better to to never have
54:06
people accept misinformation in the 1st place.
54:08
And pre bunking works kind of like like an inoculation where
54:11
we give people a weakened instead of just throwing facts
54:14
at people. It's a weakened dose of the
54:18
misinformation or the techniques used to produce misinformation.
54:21
And then you deconstruct those in advance for people.
54:24
So to build up cognitive immunity or, or mental
54:26
antibodies. And interestingly, initially we
54:29
did that with the scientific consensus on climate change,
54:32
where we told people that 97% of scientists at the time, sorry,
54:36
Marxism. Yeah, 99.
54:37
Excuse me. Yeah.
54:39
Not misinformation yet at that time I believe was before your
54:42
your paper, but the industry lobby groups used techniques to
54:45
try to sow down. So we tried to explain the, the
54:47
technique that's used by the fossil fuel industry to sow down
54:51
and so on using a weakened dose example.
54:53
Like we show them the this global warming Oregon petition,
54:57
which fakes a template of the National Academy of Science and
55:00
then has signatories like the Spice Girls and Doctor Jerry
55:03
Halliwell and whatnot and Charles Darwin and all kinds of
55:06
nonsense. So you, you pre bunk it, right?
55:08
And then we let people browse the Internet and we found them
55:10
become more immune. And we, we, we then scaled that
55:13
up in, in our interventions, using more of the underlying
55:16
techniques. Also to not have to get at this
55:19
issue of what is truth and what is misinformation.
55:21
It was more about what are the, the general tactics.
55:23
So for example, we have a new game called Bad Vax, which is
55:26
inspired by people like RFK Junior.
55:29
Bad Vax puts people into a social media simulation of what
55:32
it's like to interact with grifters on, on public health.
55:36
And you meet characters like Anne Mcdotal.
55:39
And so Anne Mcdotal loves to float emotional anecdotes to
55:42
scare you about vaccines. Then there's a doctor Forge
55:45
who's a pseudo scientist and there is a, there's a conspiracy
55:49
theorist and you, you experience weakened doses of what it's like
55:53
to be kind of manipulated online.
55:57
So you can then deconstruct those techniques and build
55:59
resistance against them. And funnily enough, this was
56:02
funded by the CDC prior to the current administration.
56:05
But you know, we've done some of these campaigns before with The
56:08
Who or other organizations. And so we can reach hundreds of
56:13
millions of people with the right partner to, to try to
56:16
prebunk. And that's what I often tell
56:18
people, the best thing you can do is, is pre bunk the playbook
56:22
of the authoritarian regime of disinformation actors that talk
56:26
about their incentives, their motives.
56:28
And that's what we do in a lot of our experiments too, is that
56:30
we just expose the motives, the incentives and the, and the, and
56:33
the techniques and the rhetorical devices and the
56:36
playbooks. And John Cook has a game called
56:39
Cranky Uncle, which does that in the context of, of climate
56:41
change. We've done videos and not
56:44
everyone wants to play a game. So we've done videos.
56:46
We've done campaigns on on Instagram with youth, for
56:48
example. Shocking actually.
56:50
So we did a campaign on Instagram about emotional
56:52
manipulation. So I've come up with this fake
56:54
campaign that yoga causes cancer.
56:56
Sounded like something RFK might say.
56:58
So baseline discernment of is this what technique is being
57:03
used here in terms of emotional manipulation?
57:05
There were words like horrifying cancer and all these things.
57:08
About 38% of people recognize it without an intervention on
57:12
social media. I mean, that's maybe some people
57:14
are not surprised, but to me, such an obvious example,
57:18
actually, I wouldn't expect many people to get that wrong.
57:20
But we learned that actually a lot of young people on social
57:23
media aren't prepared to think about what is emotional
57:25
manipulation. And so then we gave them the pre
57:27
marking intervention, which was kind of a weakened dose example
57:31
of how fear mongering is used to scare people on issues.
57:34
And then after it was something like 65%, which is still not
57:38
100%, but but you know it was, we're talking about a 2020 528%
57:42
sort of jump depending on the on the condition which is, which is
57:45
something. 55% isn't that much. I mean the anti vaxxers would be
57:48
all over you if this was a real real health intervention.
57:52
Yeah, if it was. But anyway, that's a 10 second
57:55
ad that's very cheap to administer.
57:56
So we we've come up with a book for NGOs, practitioners, how to
58:00
implement these campaigns, how to help people recognize
58:02
manipulation online. And that's what I'm still trying
58:05
to do. You know it's an uphill battle,
58:07
but how do? People find this.
58:08
Yeah, so we have a free website called Inoculation dot science.
58:11
You can find the videos games in the handbook, all for free on
58:14
Inoculation dot Science. OK.
58:16
I think well then Sandra, thanks so much for your time and your
58:20
insights on. I mean, it feels to me like, I
58:23
mean, I think we're on the same side on this.
58:24
It feels to me like we're currently losing.
58:26
But maybe the cycles will change and the wheels will turn and
58:30
we'll find that the current weird political moment of
58:33
populism and authoritarianism and misinformation does shift.
58:36
I hope so. And I hope that the way that
58:38
you're doing can can contribute to that at least to some extent.
58:40
So thanks again for everything you do, and thanks for joining
58:42
us on saving the world from bad ideas.
58:44
My pleasure. Thanks so much for listening to
58:50
Saving the World from Bad Ideas, a We Planet podcast we plan.
58:53
It's a movement of citizens and scientists dedicated to
58:56
defending the science and solutions we need to save the
58:58
world. Please do subscribe and share a
59:00
nice review. It helps us spread the message.
59:02
We'd also love to hear your feedback and suggestions at our
59:04
e-mail podcast at weplanet.org. Please also join our podcast
59:09
mailing list and we'll let you know when a new episode drops.
59:11
Just go to Saving the World from badideas.org or
59:14
weplanet.org/podcast right now. See you next time.