BBC Sounds. Music, radio, podcasts. This is a Thinking Allowed podcast from the BBC and for more details and much, much more about Thinking Allowed, go to our website at bbc.co.uk. Hello. History at my Christian brother's school was taught in much the same way.
as the Catholic Catechism. Whereas in our RI classes, we learned that the six sins against chastity were lust, masturbation, fornication, pornography, prostitution and rape. So in our history classes, we were informed that the six causes of... the French Revolution, where social inequality, the tax burden on the Third Estate, the rise of the bourgeoisie, the writings of the philosophers, the monetary crisis, and poor harvests. There was, all in all, an implicit assumption that his...
Historical facts have the same eternal veracity as God's commandments, but a new book argues that the Internet has destroyed any such claims to historical certainty. It's appropriately titled, History Disrupted. How Social Media and the World Wide Web Have Changed the Past. And his author is Jason Steinhauer, a global fellow at the Wilson Centre, a non-partisan Washington think tank. He now joins me. Jason! Let's talk first of all about the purpose of your study. What are you setting out to do?
So I'm interested in answering a question. And that question is, what have the web and social media done to our understandings of history? How has these new technologies and these new platforms changed, altered? how we understand our past and i got interested in this because there's so much discourse and so much talk about how social media and the web have changed
so many other aspects of our lives, whether it's how social media has affected journalism or how it's affected our politics or even how we date and live our lives. But no one had looked at how the web and social media have changed what we know about our past. Let me come in here with a clip. I've just got a little clip here. This is from a 2013 BBC4 documentary called The Last Days of Anne Boleyn. On the 19th of May, 1536...
One of the most infamous episodes in English history moved towards its gruesome conclusion. Anne Boleyn, Queen... and second wife of King Henry VIII was taken from her quarters in the Tower of London and with the single blow of a sword became the first Queen in Britain's history to be executed.
To Jesus Christ I commend my spirit. Lord God have pity on my soul. Now Jason, that was a series, a BBC series, which featured... narration had some re-enactments it had a whole number of expert opinions from historians and novelists now that sort of factual program really has been the traditional way in which people have learned about the past after school and after university
But what do we know now about the main sources that the public now relies on in terms of how they gain an understanding of history?
yeah well the way the public comes to understand its past is a very complex process and it starts even when we're very young right we have family histories that we learn about we learn about histories of our community we see monuments and statues around us we see old buildings around us we learn in school we learn through television programs or radio programs which is the one you just aired
And increasingly these days, we're now also encountering historical information on the web and social media sites like Wikipedia, Instagram, even TikTok. And what interested me and surprised me when I talked with people for this book, I did a lot of interviews, was how often those platforms came up as sources for historical information for people. And Wikipedia probably... sort of self-evident that that's a place people will get historical information from.
But, you know, I would talk to a young journalist, for example, and she would tell me that she gets her historical information from Instagram. Or I would talk to students who are in high school and they would say that they were watching these videos about history on TikTok. And so that. reaffirmed for me that there was something going on here that we needed to better understand and wrap our minds around
I mean, many people listening to this will think, but the internet is full of history videos, history blogs, history podcasts, social media accounts, historically informed news articles. They're all there. But you want to suggest that this excessive information about history...
It's not necessarily positive. I mean, why not? It turns out when there's this massive expansion and massive explosion of information in front of you, it can often become very difficult to figure out what to trust and what is accurate.
You might see one video that's by the BBC. You might see a next video that's by a professional historian. Then the next video you see might be by a conspiracy theorist. Next one might be by a white supremacist. Next one might be a hobbyist in Australia. Next one might be a... a set of teenagers.
And ultimately, the onus is on you as the user to figure out which ones are authoritative, which ones have accurate information. So more information does not always equal a better understanding. Another perspective which you bring to bear upon. this is the way in which historical or history online comes to our attention, you know, what it is that brings us to our attention. Tell me what are the factors at work here?
One of the things I was asking in this book and trying to figure out was why do we see certain pieces of historical information online and not see others? and it turned out the more that a piece of history content online and i call these e-history so the more that a piece of e-history
aligns with the values of the social web, the more likely you are to see it. In other words, the social web itself privileges certain types of information, just based on the way it's designed and the way the algorithms are formulated. And the more that a piece of e-history sort of conforms or aligns to what the social web prioritizes and privileges,
the more likely it is to show up in your newsfeed, regardless of how accurate it is. So accuracy and rigor and sophistication and depth of knowledge have really no bearing on whether a piece of historical information online will come to your attention. Some of these conflicts around e-history have come more fully into public view and you attribute that to the rise of Twitter in the late 2000s and early 2010s.
Tell me a little bit about the Twitter account, which has the handle at History in Picks, and what do you think this epitomises? So yeah, History and Pix is an interesting example. This is an account that was run by two teenagers in Australia. And the idea was to basically send history content viral. So I say in the book that it epitomizes something called a viral past.
and basically by sending information viral through networks it would garner attention it would garner influence and it would garner funding and that's exactly what happened they actually built a company around this idea that raised millions of dollars If you create content that goes viral on Twitter, regardless of whether it's accurate or not, it can lead to fame, attention, money, and influence. And that is exactly what happened in the case of History and Picks.
Now, the clash there came from professional historians pushing back against an account like History and Picks and saying, wait a minute, why is the ecosystem rewarding this type of behavior, this type of content? Why is the fame and the...
and the attention going to this online viral history site and not to professional scholars for example and those clashes played out very publicly on twitter in the 2010s around this account but it's indicative of how the social web was not built to prioritize accurate
historical scholarship. It was built to prioritize signals of attention. And what History in Picts was able to do was generate a lot of attention. It led to it having four million followers raising money and getting lots of influence this is not just true for history and pics it's also true for disinformation campaigns disinformation campaigns that leverage history and leverage the mechanics of the web and put history into the web in a that the web wants to see it.
they can actually have tremendous influence i talk about a an app called time hop and time hop was an app that took an element of your personal past and put it back in front of you in your news feed as a way to get you to click And we're all familiar with this. We've seen it. If you've ever used Facebook or any of these sites, there are similar ones on other sites where you get a memory from your past or, you know, on this day you were here. So it gives you this little micro dose of nostalgia.
your feed that you then click on and that is a way that it generates income and money and revenue for the company. timehop did exceptionally well on facebook it became a very popular app and so facebook realized that actually this idea could become the centerpiece the linchpin so they then actually borrowed the idea from TimeHop to create this sort of on this day feature in order to get us to engage with the newsfeed more, to click more, to scroll more, to stay on the site more, to get
advertisers to pay more for ads and to make the platform more economically viable and that's exactly what happened. And also to influence politics as you talk about the way in which this technologically induced nostalgia in a more general sense central to the politics of the 2010s. Yeah. You think about someone like Donald Trump in the United States and the idea of make America great again.
That very concept, that very phrase has nostalgia embedded in it. Even the Brexit campaign had a lot of nostalgia embedded in it. And so it made sense then to think about how these political movements that had so much nostalgia at their...
leveraged technologies like Facebook that were actually predicated on digital nostalgia, right? And so it makes sense when you combine A Trump campaign, for example, that is premised on this idea of a nostalgic return to American greatness with a platform like Facebook. which is privileging nostalgia in the feed and using nostalgia as a way to keep people on the platform, when you marry those two together, you get a very potent political combination.
Let's turn to Twitter's growth, which began to stall in the mid-2010s, and in its place came Instagram, which surpassed, I think, 500 million users in 2016. Let's talk about Instagram's part in disseminating historical knowledge. The more I looked into it, the more I realized that there were some actually really large accounts on Instagram with millions of followers that trafficked in history content, including...
an account called history cool kids which i was not familiar with on all these platforms there is some good historical information there is some bad historical information and there is information that's sort of in between but instagram is a visual medium And so on Instagram, if you're going to do history, you have to lead with visually arresting imagery. And a lot of times it turns out on Instagram that turns that winds up being war photographs or photographs from big dramatic conflict.
where there is lots of imagery to pull from. So you see on Instagram, there are a lot of history accounts that use World War II imagery, sometimes World War I imagery, sometimes Holocaust imagery that become very popular because they have this large cache of dramatic photographs that they can pull from, and they use those photographs to populate the feed, to get visitors to stop from scrolling, and as a way to get people's attention.
You also write about the way in which journalists only use history if it's in response to current headlines. Expand on that for me. Journalists obviously have an interest in history. Many journalists love history. But how does history actually wind up in the news media published on a website or in a podcast or on a radio interview like this? In order to fit into the news cycle and the news apparatus, the history has to have a news hook.
That then plays a big role in what histories get seen. For example, because social media is such a newsworthy, buzzworthy topic, my book is getting more attention than maybe a book that is about the Ottoman Empire in 1920. 10 to 1912. So that is one of the things that I think information consumers need to just be aware of is to recognize that the histories that we see in front of us, whether it be in the news media or on social media,
have reasons that they are selected. What effect will the way in which history is now being disseminated have upon our future lives, our future understanding of the past? One thing we're seeing throughout this period of the social web and history's explosion on it has actually been a hollowing out of the history profession. If you look in the United States context, the number of history departments is declining.
of history professors is declining, and the amount of funding for history museums and history organizations is declining.
Similar to what happened to music when suddenly music propagated online for free in the early 2000s, late 90s, early 2000s. There's so much history content available for free on the web today that it sort of... uh begs the question of why anybody should pay for it and i heard this again and again and again from people that i interviewed so i think this is one of the existential questions that is facing not just profession of history but all of society
Do we value professional history, professional scholarship enough that we're willing to invest in it and pay for it and continue to support its existence? Jason Steinhardt, thank you very much. Well, I can't really leave the subject of history and its interpretation without a favourite quotation from that wonderful Child's Guide to History, 1066 and all that. Here is the author of that W.C. seller on the origins of America.
One day, when George III was insane, he heard that the Americans never had afternoon tea. This made him very obstinate, and he invited them all to a compulsory tea party. at Boston. The Americans, however, started pouring the tea into Boston Harbour and went on pouring things into Boston Harbour until they were quite independent, thus causing the United States. Now, as usual, I'd love your comments on all such matters, fanciful or factual, at thinkingaloud at bbc.co.uk
Now, this has been rather a sad time for me. I learned quite recently of the death of Victor Lewis Smith, the satirical writer and television director. who'd been my, well, my scurrilous friend since we first met over 40 years ago when he was a music student at York University and I was a junior lecturer in sociology. It was enough of a shock to send me scurrying to medical sites on the internet in search of information about, well, my own prospects.
And I dutifully filled in all the boxes. How much did I drink? How often did I exercise? Did anyone in my family suffer from cardiac problems? According to one such online survey, I should have passed away already. I was living on borrowed time. Well, this online determination was, of course, created by the use of algorithms, the use of existing data to predict the future.
Well this can be a controversial matter. In 2020, thousands of students in England were angry about the use of an algorithm to determine that year's GCSE and A-level results. Students were unable to sit exams because of the lockdown, and the algorithm used predicted grades from teachers and the past performance of each school.
Tens of thousands of A-level and GCSE pupils in England will, after all, receive grades decided by teachers' assessments rather than by a controversial algorithm, following fierce criticism from parents, teachers... The Education Secretary has apologised for the distress caused to students. Labour said ministers had been forced into a screeching U-turn.
Well, as we just heard in that clip from the 6 o'clock news here on Radio 4 back in August 2020, the government reversed their decision and relied on teacher assessments instead. One of the major concerns was that the use of algorithms affected the most economically disadvantaged pupils. Just one example of the potentially negative impact of using algorithms to predict the future.
an issue explored by my next guest. She's Helga Novotny, Professor Emerita of Social Studies of Science at ETH Zurich and author of In AI We Trust. power, illusion and control of predictive algorithms. Helga... There are many examples of algorithms making small as well as big decisions about our future lives without us necessarily knowing how or when they do it. Give me some more examples from the trivial to the more concerning.
Let's start with the trivial examples. Whenever we make a click in order to buy something or to make a reservation in a restaurant, we want to get information. where to spend our vacation, which book to buy. We confer our data to the provider, to the big corporations, and these feed the kind of pools of data. that are necessary in order for the predictive algorithm to start to extrapolate from the past what they can predict in the future now these algorithms are invisible to us we don't even
think about it. But as you cited this very striking example that happened during the lockdown, which, by the way, I mentioned in my book. It shows that there are very harmful consequences that may be connected to the use of predictive algorithms. And in the US, for instance, the courts and the police. use them in a routine fashion. to help the judges to make decisions about who will be granted bail or who has to go to jail right away, which person to be tried is likely to be a recidivist.
and then the sentence will be harsher than if this is not the case. Now, the problem with this is that if you have no human intervening at the end to see what the algorithm proposes, for decision-making or in fact makes the decision, then we are in deep trouble. In the end, there has to be some kind of human oversight. The data come already with a bias. We all have biases as persons, but given this enormous amount of data.
All our personal biases are part of the data. The racist bias is a very strong one. Also Jason mentioned how history is presented. You have also biases there. that enter into the data and there is no one to raise the flag and say, wait a moment, this is not accurate, this is wrong, and we have to stop this. And you enter another concern about algorithms. People then seem to be aware that predictive algorithms can act really as...
Self-fulfilling prophecies. What do you mean by that exactly? How does that work? The phrase was coined by an American sociologist way back in the 1930s, and his definition was... If men, today we would say men and women, define a situation as real, it is real in its consequences. In other words, if people perceive a situation or a prediction in this case and believe that the prediction will become true.
It actually means that they will change their behavior in order to make it become true. That's a self-fulfilling prophecy. And the classical example is the run on the bank. The rumor spreads that... you can no longer get your money back from the bank. So everyone starts to run to get their deposit money. and of course the bank cannot pay everyone because it does not have the means and it gives an enormous power to predictive algorithm if enough people believe
that it is true and it will actually happen. We all of us tend to wonder about the future. I was talking before about the future of history. And you suggest this isn't really a new dilemma. We find that every civilization we know of... People wanted to know what the future has in store for them. And therefore, they used divination. So you had people who were experts in divination. In China, for instance, they used...
the bones of tertoises or of sheep, which they held over the fire. Then the bones showed cracks, and they interpreted the cracks in order to tell the future. finding every culture this kind of oracles and and divination practices now today we no longer or at least the vast majority of people no longer uses divination we have science
We have foresight reports that the governments write. But nevertheless, the wish to know the future is as strong now as it always was in history. Now, one very important... difference between these thousands of years of human history and today's world. It was enabled by science and the European Enlightenment, that all of a sudden people started to realize that the future is no longer set in stone. It's not predetermined by the gods or by...
But rather, the future is an open horizon. We don't know what's going to come. And this is, in my view, an enormous discovery of humanity. But if we start to believe too much, if we start to trust what these predictive algorithms tell us about our future, we fall... again into the trap that we start to believe when, after all, the world is a predetermined world and we would relinquish one of these great experiences of humanity to some extent the future is in our hands and we can do something
to bring about the future that we would like to have. You've talked very eloquently about the danger of relying upon predictive algorithms for our assessment of... what the future is going to be like. I mean, currently we are facing ongoing crises of many kinds. I mean, there's climate changes, threats to democracy, increasing social inequalities. I mean, on the basis of your research...
How best do you think human beings can face these challenges? The algorithms do not know the future. The future remains uncertain. They allow us, in the best of all cases, to see further ahead. When it comes to these crises, We have to accept that so much of the future remains unknown, but it also offers opportunities that we do not see as yet.
And it's obvious if we want to get out of the predicaments of climate change or, you know, the fallout from the pandemic, inflation, the energy situation that we now face and much more. I mean, it forces us really to take stock, and most likely we will have to come up with another mode of how we see that our societies can and should. function better in in the future.
We live through a crisis, but we have also to gather the courage and the stamina to deal with it. The worst thing to happen is if people fear what is going to happen. because fear stifles any kind of action. You've spoken very lucidly about the danger of over-reliance upon predictive algorithms, but surely, I mean, in some cases they can make better decisions.
in terms, for example, you were just referring to protecting the environment. I mean, if we talk about energy consumption, I mean, if we have all the available data, we can really talk about the future, can't we? We can rely upon algorithms. to give us a sense of how the future is going to be? These are the positive sides of predictive algorithms. The environment is one domain. I mean, we have so many sensors.
working with algorithms that let us, you know, measure the kind of changes, the degradation of the environment, how fast the glaciers make. All this is only possible because we have these digital means of measuring. And then, of course, based on these data, you can make predictions what's going to happen in which parts of the world in the next five years. And you see it in the medical diagnosis have progressed and benefited enormously from predictive algorithms because they allow.
the machine to see patterns in a much more accurate way than any doctor. But mind you, all these predictions that come from science are always couched in probabilities. So there is no absolute certainty. that is attached to them. Power now seems to be more and more in the hands of the technology giants. I mean, presumably you want to propose some regulation in order to limit that power and to bring the human dimension back into play.
Absolutely. It's the duty of the state to regulate the technology that is harmful to its population. The problem is that in the past, regulation was always done by the nation state, by governments, for their population. now as we all know these technologies are global they are everywhere and there is no world government that can regulate them but i am really convinced that without regulation, we would hedge down a very slippery slope. Helga Novotny, thank you very much for talking to me.
Last words, which some might feel are particularly appropriate in these troubled times, from historian Michael Corder, who wrote, I mean, who can, alas? It is amazing that they defiantly ignore the past. That was a Thinking Allowed podcast from BBC Radio 4. You'll find a treasure trove of other Thinking Allowed programmes. on BBC Sounds.