Skip to content
Bullshit Detecting With Nick Enfield episode artwork

Jul 6, 2017 29 min

Bullshit Detecting With Nick Enfield

from Fourth Estate Hosted by 2SER

Nick Enfield is a professor of linguistics at the University of Sydney, and head researcher at the Post Truth Initiative. Enfield has gathered a team of researchers including data scientists and software developers to create a 'bullsh*t detector' hoping to stop the spread of fake news, lies, and propaganda. Miles Herbert sat down with Nick to discuss how we should navigate this current political climate of misinformation. Also Kid Rock.

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

Bullshit Detecting With Nick Enfield

0:00 29:20
of matches

Transcript Auto-generated, may contain errors

Your summer travels deserve an upgrade. With select Chevrolet, Buick, GMC, and Cadillac vehicles connected by OnStar, you can stream, game, and more all the go. Whether you're headed to a music festival or a cottage getaway, you can access hundreds of apps on Google Play right from your center display. Explore OnStar tech today at OnStar.ca and discover how to take your summer to the next level.

OnStar. Better never stops. This episode is brought to you by L'Oreal Group. Beauty is a powerful force that moves us.

That's why L'Oreal Group has built a business that is inclusive at its heart with 100% of its brands, championing diversity. With 25,000 professional opportunities for people under 30 worldwide and 54% of leading positions held by women, diversity is a strength that helps L'Oreal Group create the best beauty products for all people. Visit L'Oreal.com to learn more. I started today by asking Nick what a linguist does and how that might intersect with the world of journalism.

Well, I'm a professionally trained linguist, so that means that I studied linguistics at the university and got the sort of basic skills of linguistics through that. Linguistics is sometimes thought of as being the field where you learn lots and lots of languages and you become like a high-performance sort of translator, but that's not what we do. What we do is we try to study the kind of essence of human language. So in my area, I was particularly keen on kind of world travel and getting out around the world and seeing different people, and linguistics is a great way to do that because you're trying to kind of tap into different languages and different cultures.

So that's kind of what got me into it, and what I do now is kind of, you know, basically grounded in exactly that. So I do fieldwork in Laos. That's my country of specialization and sort of other parts of Southeast Asia. And what I do is collect recordings of people using their language in everyday life and really kind of examine how they use it, what they're doing when they use language, sort of looking at some basic questions of human social life and what role languages play.

How do you think the study of language intersects with journalism? Well, it's crucial. I mean, language is the tool of the journalist. Obviously, images are as well, but images are intimately kind of related to language.

So, you know, what I do is really about kind of expressing yourself and through that having an impact on other people and trying to change them in certain ways. And sometimes that's very kind of trivial. So if I say, pass me the salt, you know, in a way I'm kind of manipulating you in some sense, right? And so I can look at some very basic uses of language that make impressions on you.

Or if I tell you my views about something, you know, I thought that movie was terrible or something, I'm changing you in some very minor way by changing your kind of beliefs about me or what have you. Journalism is obviously, you know, that is kind of cranked up to 11, right? You're trying to deliver the truth and deliver the facts in a way that's relevant to people. I mean, this is my naive view of what journalism is.

But, you know, what you're really doing is using language as a tool to change things through informing people and what that fundamentally is, is changing their state of mind. So you're the head researcher at the Sydney Research Excellence Initiative, a new project examining fake news, alternative facts, lies, bullshit and propaganda, all terms that have kind of been banded around endlessly by talking heads and politicians over the last year. What we call the kind of post-truth age, you know, so the initiative, the Sydney Research Excellence Initiative is a larger thing with a range of different projects that are funded and ours is called the post-truth initiative. And so things like fake news, propaganda and so forth all come under that heading.

And it's a very good time to be studying these things because they're very topical. And, you know, there's been a huge amount of news about these problems of kind of getting the facts straight and so forth. But of course, there's nothing new about that, as you just hinted, you know, we've always had lies and propaganda and coercion and all those kinds of things. So in a way, you know, what we're really looking at is a classic age old problem.

That doesn't mean the problem shouldn't be studied, obviously, because it's still a diabolical problem. But it seems like there's something really different going on today in this domain because it's just through the ways of getting information around the globe and getting information to other people. I mean, you know, it's been a revolution, revolution after revolution over kind of recent years in terms of the flow of information. So there are new ways to kind of spread bullshit and propaganda and that kind of thing.

And there are sort of new speeds at which one can do it and new kind of cycles. So there's a sense of urgency. Some people kind of have the view that, you know, we shouldn't be alarmed about current state of affairs. But I think that, well, we may not need to be alarmed, but we certainly should be concerned and try to kind of work on the problem.

I mean, there's a cultural problem, obviously, where people are ready to take on ideas that they hear without checking them, without looking at the evidence behind them. People are willing to trust certain sources without actually having great reason for trusting them. And obviously, that's a serious problem, right, when it comes down to things that matter, you know, fundamentally, for example, being things like climate change and so forth. If you make the wrong decisions because you have false beliefs or because you don't care what the truth is, then, you know, ultimately that's suicide.

Right. And so that name, the Post-Truth Initiative, you know, in response to things like Brexit in the UK and Donald Trump in America and kind of the rise of populism across the world, that word post-truth is now in the dictionary, you know, it's now a recognized word. Do you think that this post-truth initiative and the research you guys are doing was an indirect response to things like Brexit and Donald Trump? Yeah, it was.

I mean, the idea of having a kind of a large group of people working together on this was certainly triggered directly by Brexit and Trump in terms of the timing. We mentioned those things sort of in the materials that we used when we were proposing kind of doing this. You know, as I just said earlier, the question of sort of how new those things are, you know, it's pretty clear that they're not especially new. We could have had an initiative like this earlier on, you know, around other kinds of phenomena.

But, you know, there's something really galvanizing about those two stories, right, the Brexit and the Trump, just because they were so shocking to so many people. Maybe some people weren't that shocked, but it's only in my bubble. And me personally, I was quite shocked by those two developments. I just wouldn't have imagined that either of those things would have happened if you'd asked me the day before.

So, you know, I think that a lot of the people I work with were probably in the same boat. And one of the reactions to that is, you know, obviously I'm missing something here. So we better do some research and try to figure out what's going on. One of the aspects of this initiative is what you coined a bullshit detector with the ability to detect stories or so-called fake news that are avoided any rational thought, arguments or truth.

What is this bullshit detector and how does it work? So this is something that we've started developing in collaboration with computer scientists at the university. And it's really a way of trying to explore some of the new developments in, well, especially in sort of voice recognition, but also in data processing and kind of online way. I mean, we started out the proposal with a kind of a dream concept that you could sort of crank up the current sort of technologies that were allowing us to do things like fact check and that sort of thing and use it as a type of real-time kind of fact checker in a sense.

So that's what we sort of set out to do. I mean, there's some pretty amazing new kind of voice recognition technology where you can do searches in, you know, in the same way that we used to do search in text, you can do it just, you know, if someone's given a speech and you have a video of it, you can search for sort of phrases that they've been using and so on. So that kind of thing made us really want to look into the possibility of triangulating statements against established evidence that we have on those same statements and so forth. So, you know, the dream kind of device would be one that would sort of set off an alarm and say this is false and so forth.

Obviously, in terms of sort of implementation, it's a really hard thing to do. So we've been working pretty hard on kind of narrowing down the set of things that we could do to do something that's tractable and it has some kind of leg. So right now we've been developing a way of doing this that focuses on individual people. So, you know, you can look at a public figure and you can take the statements that they're saying and compare them to statements that they've said in the past about the same thing and look at the kind of, give them a sense of kind of score of consistency in terms of sort of what they've said.

Now, obviously, people can change their minds about things and that's fine and, you know, new evidence can come to light and we can have a new stance on something. With politicians, you know, we don't generally see that being a pattern that it's kind of new evidence that makes them change what they're saying. It tends to be something about the context that they're in or who they're talking to and so forth. So at the moment, kind of what we're really exploring and still, you know, in the development phase is this kind of measure of consistency for certain people in terms of their sort of history of public statements about a topic.

And that really kind of looks at, in a sense, gives us a way to rate them to the degree to which they're a bullshit artist. So, I mean, this idea of bullshit has a kind of technical definition, if you like. In philosophy, there's a sort of distinction between lying and bullshit. Lying being saying something that's untrue, knowing that it's untrue, with intention to deceive, right?

Bullshit, on the other hand, it's obviously similar, but the idea behind bullshit, as we see it, is that you are saying stuff and you don't care whether it's true or false. You're just saying stuff that you think people want to hear or that's going to impress people that's going to let you get your way. So you actually don't care whether it's true or false. So one of the kind of effects of being a bullshit artist is that one day you can say one thing and the next day you can say another and then you can go back to the first thing and the day after that, right?

So that kind of, at the moment, is how we're trying to get away into sort of really trying to detect partly people's sincerity, partly people's kind of manipulativeness, all of those kind of signals of being a kind of propaganda artist. Besides comparing and contrasting public statements past and present, how else would this bullshit detector determine what rational thought is? How would it look at text or analyze speech to determine its merit? Well, rational thought is, there's obviously a lot to be said about rational thought and what it is.

Logic is kind of at the heart of it in a way. I mean, there's certainly a major tradition that looks at kind of logic as being at the basis of rational thought, but fundamentally without even needing to kind of go into technicalities of how logic works, essentially rational thought is a way of coming to conclusions based on knowledge and evidence that you've been given beforehand. You know, being a rational creature means that you can arrive at new knowledge that nobody had before and you can arrive at that knowledge through learning things and basically putting two and two together, right? So, you know, there's a couple of elements to that that are really important to make sure you've got straight.

So one of them is the processes of inference and deduction and does, have you actually added two and two to get two or have you multiplied them? You've got to get your kind of processes of inference right and know what you're doing. And the other part is, you know, what are the facts that you're starting from? Okay, sort of is this a garbage in garbage out type of thing?

You could operate perfectly logical kind of inferences and come to conclusions but if your facts are wrong that you started out with, you might have a valid argument sort of on internal ground but it would be a useless one because the conclusion is actually based on kind of falsehood. So when we talk about rational discourse, we're really talking about the kind of discourse you see in public life where we have to make decisions that affect groups, okay? So a lot of times we're trying to figure out what's the right thing to do on behalf of a group so to be fair and to do things that make sense. You really need to have confidence in the kind of premises that you're starting with and all of that really boils down to this problem of, you know, knowing what's true, what's not true, knowing what's a fact and what's not a fact and that's where, you know, in this 24-hour news cycle in the world of social media, news travels really fast.

I feel like politicians are always saying something that's always story-breaking. So how would this bullshit detector work to combat an onslaught of information, you know, information overload? Well, I wish it could do that. You know, I think that a device like this in a way that we conceptualize it is not really designed to do that.

I mean, it's designed, you know, you present it with some information or with a kind of corpus or with a set of recordings or what have you, you know, and if it's working, it will tell you something about the facts or the coherence of it or the consistency of it. I could imagine in the future it could sort of feed into some kind of filter, you know, or allow you to set some kind of threshold. If you could trust the device, then you'd set it and it would actually stop bullshit coming into your inbox, if you know what I mean. That would be great.

That would be a nice thing. Of course, you'd have to trust it quite a bit. And this is, you know, there are obvious dangers to that. So we're already in these incredible bubbles and things are getting filtered left, right, and center through, you know, all of the kind of advertising-based filters that are derived by algorithms, you know, in very clever ways based on kind of all the input that we give to these companies and so forth.

So, you know, certainly the potential's there for doing that. Again, it comes down to who's controlling that technology, how you're setting the parameters to kind of edit out all the bullshit. But I think it would be a great thing if it's something much better to have it filtering out the bullshit rather than feeding more and more of it in as is currently the case. Yeah, I wanted to ask you, you know, even the mainstream media now is so heavily mediated.

It's really common for citizens to only get their news from one source, whether it's the Murdoch Press, Fairfax, or, you know, just friends on Facebook. But the information people receive is so heavily mediated, is there a chance that this bullshit detector might just be another filter ensuring people only get the information they really want to hear? Well, you know, you've kind of hit the nail on the head of a certain type of problem. There could be a chance, you know, in the future if technology like this were kind of applied and scaled up, you know, I can imagine, yeah, that would be one of the kind of problems.

So that would be kind of a question of implementation, you know, what uses the kind of tool was really put to. So I don't think it's inherently a danger, but, you know, in any kind of technology, at least they can be used for good or bad or they can be used well or not well. So, you know, I think the potential is there, but it's certainly not a necessary, it wouldn't be a necessary consequence of that type of a device simply because, you know, everything we're working with in the information world, you know, is ultimately a kind of seerful filter. And so the real question is, well, which ones are the good ones and how do we apply them well?

So let's talk about social media. Platforms like Facebook and Twitter, even though they say they're not media companies, are increasingly the only place people get their news. How would this bullshit detector be able to assess information on social media platforms? I'm not sure being on a different platform would really change the way that something like a bullshit detector would do its work because essentially, if it's going to work, what it's got to do is to take a statement or an assertion and evaluate the kind of, you know, either the truth of it or the consistency of it in relation to that source.

So, you know, I think this is exactly why things like Twitter and Facebook work for people as sources of news is that, you know, they're kind of indistinguishable from the news outlets in that, you know, you're getting assertions being made about states of affairs. You know, there's a place and there's an incident there, you know, a rocket goes off or, you know, a bomb gets dropped or what have you. And you read about it and then really for something like a bullshit detector, it's just, it doesn't really care, you know, as a kind of computational device, it doesn't really care what the story is about. It just wants to check it against evidence that it has.

So I guess you would like to see, and there's certainly a very strong kind of public demand for this. You'd like to see some device like that being used wisely with that type of media. And it's, you know, it's very, obviously very dangerous just because people are able to manipulate it in whatever ways. I mean, I should hasten to say that a lot of people would say, well, it's never been any different with not social media, with traditional media.

That is, people are manipulating it left, right and center for their own kind of devices. It's just that it's now splinter and splinter and splinter, you know, and there's just so many more bubbles than there ever used to be. Yeah, so I think it's come out, you know, pretty recently that Facebook does monitor its content and they do make editorial decisions about what content ends up on their pages and what you see in your newsfeed. I guess my question would be, could we see, you know, like a mobile app plug-in, you could plug into your social media platforms that as content was coming up on your newsfeed, it would detect it come out at you in flash and be like, fake news, fake news, something like that.

Yeah, I could imagine something like that and I think things like this exist already that there's some sort of warning comes up and it says that, you know, this may be fake just because it detects certain sources get blacklisted and so, you know, it's a pretty basic kind of brute form of filtering. It doesn't actually cut the story out but it just tells you read this with caution. So that would be a very sort of coarse filter that would capture some kind of things and you would hope with more sophisticated devices like the one we're trying to develop at the University of Sydney it could be something much more sophisticated so the machine wouldn't care where the information is coming from it would just evaluate it against the available data that it has. So I think that would be an obvious kind of application it would be a nice one to have but it would certainly have to go through a long sort of phase of development because it's pretty annoying having things flashing going off, you know, in your newsfeed and stuff.

I mean, you don't really want that. You just want stuff that you can trust. But then, of course, if there's a heavy filter in place there, you also don't want to miss out on stuff that's relevant or also true. Well, so let's track your story.

Do you have a favorite fake news story you've seen over the last year? Favorite fake news story? Gee, you know, actually, I kind of ignore them and I haven't really tracked any of them. On our website, the Post Truth and Issue, we actually have a feed on the front page that, you know, that we get, you know, all the worst rated things on the fact check sites that they kind of come up and I see them, you know, the last one I saw yesterday was Death Hoax.

Kid Rock died, I think, fakely. Right, so let's say Kid Rock dies fakely. That starts probably, you know, in some corner of the internet, somewhere like 4chan or Reddit, places where groups like Alt-Right and pro-popular trolls hang out and kind of just generate content and fake news, whether it's to influence political parties or just trick people on the internet about quote-unquote famous celebrities. And then the story is perpetuated on social media platforms and then in some cases pops up in traditional news outlets.

So where does the bullshit detector come in if it's only at the very end of it once it's in the mainstream media, if it's on your social media platform? Is that maybe too late to stop this information from infiltrating society and people's public discourse? That's a really good point and it's certainly a concern. So one of my colleagues, David Schlossberg, who's in the Sydney Environment Institute, is quite so strong on his point and he's very much more concerned not only with kind of knowing something's bullshit when it occurs but actually finding the source of it, right?

So, you know, just what you described, you know, there's a question of accountability for these kinds of stories and who spread them or who started them. Who was that? It was just in the last two days. What story am I thinking about where?

Oh, it's the CNN head. This is not fake news but it's the guy who did the video with the CNN logo on the head of the... Someone tweeted a photo of Trump while he was in the WWE years ago and someone used Photoshop to put CNN on his head and then tweeted it out and then it somehow got to Donald Trump who also retweeted that. Yeah, so they had actually found who was the person who originally made that video and it was on Reddit or whatever.

I don't know if you've seen this but this person was kind of talking and speaking out and saying he was kind of ashamed of what he'd done and he was just doing it to sort of fire people up and this kind of thing. I mean, anyway, so I digress. I was just trying to think of it. So what's the source of the bullshit?

And for a certain story like, you know, Kid Rock died, it's not going to make any difference to anyone who knows why someone wants to do it. They just want to kind of see their story spread or something like that just for fun. But it's much more sinister when you're looking at stories like, you know, there's such a thing as clean coal. Is there such a thing as clean coal?

Well, you know, a lot of people I talk to, for example, at the Sydney Environment Institute say it doesn't make any sense. It's an impossible concept and yet it's out there. So the question is, you know, who devised this idea, that particular phrase, you know, was it devised in a boardroom somewhere? Was it sort of released like a virus?

You know, all of that is thinkable and, you know, that's a really interesting question is to sort of try to track this stuff. So that is a secondary question really to sort of detecting bullshit is actually finding the source of it and it's a hard problem. I mean, I think there are ways in which you can do it if you sort of, if you're an investigative journalist, essentially, you can try to kind of track those things back and find out. But how to automate that sort of thing, you know, is a tough one.

When you say it's too late, I guess, yes, no. I mean, it's better to be late than never in this domain, I think. But certainly once stuff has got out there, it's very, very hard to stop it. And it's also very hard to stop it.

Human psychology kind of plays into all of this. So, yeah, I think you're right to say, you know, maybe it's too late, but yeah, I'd say better late than never. You just talked a little bit about psychology. Earlier we talked about data science.

It sounds like you've assembled a bit of a team over there at the Post-Tooth Initiative. I see you have research in physics, philosophy, data science, international relations, media and communications, as well as software engineering. It seems to me traditional forms of media have had a rough time combating quote-unquote fake news. Do you think an interdisciplinary team of researchers are more equipped to get to the root of this problem than journalists?

I think teamwork is always a good idea. You know, I think the journalists can learn from academic scientists, sort of social activists, anybody who's kind of related in any way to what they do and vice versa, right? You know, I think in my own experience as a researcher, it's always a good idea to collaborate. It's always a good idea to do interdisciplinary work.

And it just helps you to think outside the box a little bit and pull together different ways of thinking to sort of see, you know, potentially new solutions I hope. Well, for example, in America, Donald Trump, he's now the president of the United States. So traditionally, if I'm a journalist and the president of the United States is saying something, that means I should cover it. In Australia, we have Pauline Hanson, you know, and she's a senator.

And traditionally, as a journalist, if a senator is saying something, that means I should cover it. But I think in this post-truth world, that might not actually be the case. As an academic who is examining this world of misinformation, should journalists be moving away from traditional models of coverage? Well, I guess a number of things there, you know, if you think that it's bullshit, therefore we should leave it alone.

I think there's something you said for that, and certainly in terms of things like, you know, Hanson Trump, I'm getting pretty tired of all the media that I see about these people because, you know, you can repeatedly see that it's bullshit. So it's not that it's irrelevant to our lives or, you know, uninteresting or things like that. But, you know, it's over the top, I think, the amount of kind of attention that's being paid to it. So, you know, there's an obvious reason, I guess, and it's the one that many people would point out, and that is that the kind of elephant in the room about lots of journalism is that it's there to sell papers and, you know, there to attract advertising money on programs that get high ratings and stuff.

So whatever is more attractive or at the very most basic level, whatever grabs people's attention is going to be what journalists will cover. What they should cover might be another kind of question. Well, before I let you go, I just have one more question. If the bullshit detector isn't fully operational, information to check the source and check if you can get multiple sources that are making the same statement, I mean, it's a pretty basic aspect of internet literacy, you might call it, but I would say if you wanted to be responsible about how you handle information, whether it's a daily telegraph or something that you really trust, if you're going to make a decision based on it or if you're going to pass that information on, then the best thing you can do is try to find at least one, preferably two or three independent sources of the same kind of piece of news.

But with respect to facts, the fact that we've got this incredibly powerful information-gathering devices in our pockets means that we kind of have no excuse for just believing something when we see it. Dr. McKenfield, thank you very much for being on 4th Estate. Thank you very much for having me.

That's it from us on 4th Estate. Don't forget, you can subscribe to the 4th Estate Podcast via iTunes or your favorite download app. I'll catch you guys at the same time next week.

Frequently asked questions

How long is this episode of Fourth Estate?

This episode is 29 minutes long.

When was this Fourth Estate episode published?

This episode was published on July 6, 2017.

Is there a transcript available for this episode?

Yes, a full transcript is available for this episode. You can read the complete transcript on the episode page.

Can I download this Fourth Estate episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.