this is bbc news, the headlines: the us government is to fund thousands of hospital beds for coronavirus patients in the hot spot states of new york, california and washington. meanwhile partisan disputes in the senate have been holding up the approval of a huge stimulus bill to help an economy hit hard by the pandemic. britons have been warned to take advice on social distancing seriously or stricter measures could be imposed. prime minister borisjohnson said a lockdown like those seen in other european countries was being considered and said people ignoring advice were putting lives at risk. for the first time there's been an acknowledgement from the japanese prime minister that the tokyo olympic games may have to be postponed because of coronavirus. shinzo abe told parliament a delay might have to be considered if safety could not be guaranteed. world athletics, as well as the canadian and australian olympic committees, have said the games cannot go ahead in july.

now on bbc news, it's hardtalk with stephen sackur. welcome to hardtalk. i'm stephen sackur. the worldwide spread of coronavirus and its significant negative impact on the global economy represents a powerful illustration of the perils of forecasting. countries in lockdown, financial markets in turmoil — this isn't the way 2020 was supposed to pan out. an opportune moment then for my guest today. acclaimed writer and businesswoman margaret heffernan, to publish a book challenging the fetishisation of forecasting. is it time to embrace uncertainty?

margaret heffernan, welcome to hardtalk. thank you. i think we would all like to believe that as human knowledge expands, particularly with technology, our ability to forecast the future accurately improves. is it your contention that that is not true? it isn't my contention that it isn't true. definitely, our capacity to forecast many things has improved, specifically the weather — although it may not often feel that way. but i think what has happened is that our expectation of that improvement was continuously possible, that it would eventually end up with perfect forecastability. i think that has not

been proved and in fact, you know, both physicists and experts in forecasting will say that there is ineradicable uncertainty in human systems and natural systems, and that while we become very attracted by all of the stuff we can predict, like a gps can tell you, roughly, roughly, how long it will take you to drive home for example. it can't predict whether a cat's going to run in the road and stop everybody and hold up traffic. it can't predict if suddenly there is going to be a car crash in front of you. there are these aspects of uncertainty that are always going to be with us. sure, but probability is important. if we are to use forecasting to make big, strategic decisions about our own individual futures and our collective futures as society and as a species, then knowing what is probably going to happen and the percentage of likelihood, all of the things that go into forecasting and the results of forecasting — they are very important. they are important but i

would say two things. first of all, even professional mathematicians say that they find probability counterintuitive and hard to grasp. if it's hard for them, it's harder for the rest of us. i think the other thing is that, the real experts in forecasting say they think that with all the best they can deliver, they can be reasonably accurate about 400 days out. the rest of us, who don't have all that training and success and data and all that time to spend mulling over it, they reckon the window is about 150 days. so when it comes to many things in our lives like where to live, what kind ofjob to do, what kind of industry to be in, actually, i think we have to come to terms with the fact that life changes all of the time in ways that we can't protect and uncertainty is a feature of human existence. you say uncertainty is a feature and i suppose there could not be

a more obvious time to be discussing this than right now. you wrote this book, uncharted, well before coronavirus became the world preoccupation but here we. do you regard what is happening right around the world today with notjust the healthcare crisis being faced by populations around the world but also the economic breakdown we see happening in front of our eyes. do you see that as some sort of vindication of your proposition? i would hate to say that because it suggests that i am kind of happy about it. i'm not for a moment trying to suggest that. i'm really not happy about it but i would say that when i started working on the book about four years ago, the argument around uncertain was definitely not going my way and many people i talked to just didn't get it, just didn't buy it and thought i was flat wrong. because they felt that particularly with the rise of ai and big data... absolutely. we are going to know everything

and we are indeed masters of the universe. i think now, i would love to think that the cogent arguments in my book changed everybody‘s mind but i think the truth is, the facts have changed everybody‘s mind. i think what interests me about coronavirus is the way different governments are handling the challenge and the future challenge in very different ways. so, we have beijing where once they grasped the seriousness of the situation, they committed to, i think what one senior official described as, the elimination of the virus as a threat to the chinese public within months. they issued their forecasts of what would happen and they used central government power to deliver it. maybe they were onto something? i think what is interesting is what they were able to do was to harness a lot of resources very, very fast. i think most people saw the photographs of hospitals being built overnight. and i think in the western world, we have become so entranced by what i think of as the myth of infinite efficiency that all of those resources, we've been cutting back and cutting

back and cutting back, and efficiency is fantastic when you know exactly what you're going to need. but once you accept that there is some uncertainty in the system, you have to accept that you also need robustness, in other words, you need extra, you need slack in the system. so that when the unexpected occurs, you have the resources to throw at it. i think what's really startling about this pandemic is that we're having confront the fact that we served efficiency with such devotion that we cut that robustness out of our systems. just a really trivial example ? in all retailers and manufacturers have gone over to just in time — razor thin margins of time and money and shipping. which everybody thought was the miracle of modern industrialisation.

just in time is why no chemists have hand sanitisers at the moment because they didn't have any extra anywhere. and i think we have so fallen into this trap of efficiency thinking we've forgotten there's always going to be surprises. coming back to your idea that forecasting can take us into dangerous areas, it can be misleading and positively counter productive. a lot of people are doing forecasting today about where this pandemic is going to go, about how many people will ultimately be affected. for one, chancellor merkel in germany said that she believes up to 70% of her own german population could end up infected with coronavirus. are you saying that politicians should avoid all of that forecasting all together? i think it would be helpful to avoid that forecasting because again, she didn't attach a probability to it and it may well be that by next week, she adjusts her numbers. what good forecasters do

is they are constantly adjusting their numbers, and also, what does that mean — 70% are infected? some of those people will experience absolutely nothing, we know that kids may be infected but experience no symptoms. in other people, it'll be excruciatingly painful but not life—threatening and in some people, we know it will be life—threatening. that's a very broad spectrum which the 70% doesn't really illustrate in any kind of meaningful way at all. let us look beyond coronavirus and think deeply about what forecasting depends upon. it does depend upon increasingly access to huge amounts of data, big data is the phrase. and the analysis of the data, using algorithms, we all know now that that's the foundation stone of artificial intelligence and that's being used increasingly across governments, business ? it almost sounds to me like you're saying, don't trust artificial intelligence, it gets stuff wrong as much as it gets stuff right. it does get a lot wrong.

first of all, we know that it is biased, which is why amazonia couldn't figure out an ai to do gender neutral hiring and after two years, gave up. we know that in systems where it's been used to try and predict who will need social benefits or who's eligible for parole that it makes really gross errors. we know that it often makes, what would be if people operated the same way, illegal errors, in terms of deciding who's eligible for a job and who isn't. you're pointing to what you regard as the errors, why not look at some of the extraordinary successes of ai? you only have to look at the way social media platforms work to know that they are extraordinarily successful in targeting and reaching out and predicting behaviours. yeah, it's quite of interesting though isn't it because even although cambridge analytica made a big song and dance about how they won the election and they won the brexit vote, actually they don't know who voted how. so they claimed something they saw as a victory and then all of a sudden it wasn't

so cool to boast about it, but actually you can't prove that it was them that won it. when it comes to pandemics there's something really important here, and experts in pandemics have told me, every epidemic is different, or as they said, in the dataset, n=i— every single one is different. you can try and extract from the data general rules about what is helpful but what you know is that this is not like the last time. i understand that but again, you're pointing to the negatives and problems, but i'm seeing some of the assets the technology is bringing us, notjust ai but also genetics. the fact we're building a genetic database about all of humanity and the more we know about yours and my genetic make up, my prediliction to lung disease or yours. or my likely lifespan based on my genetic health, the more that can inform forecasts about public health strategies.

this isn'tjust about coronavirus but all forms of healthcare going forward, isn't that exciting? it is absolutely exciting and i would be the last person in the world to say that we don't want that knowledge but i think you have to keep recognising that even what genetic testing will tell us is around probabilities and not certainties. there's a huge amount that we don't understand. for example, we know that identical twins have identical dna but we also know that if one of those twins gets multiple sclerosis, if genetics were destiny, the other one would too. in fact, the likelihood of the other one getting it is just 30%. let's be honest, we still don't know what accounts for the fact that some children who were born blonde turn... there hair turns brown and some it doesn't. we don't have a genetic explanation for something as simple as that. it's not that we shouldn't keep

learning, of course we should keep learning, but we have to recognise that genetics are not destiny, that much happens to people in their lives and that if we start treating genetics as destiny, we treat as certain something which remains ambiguous. ok, so both on the abilities of ai to change our world and allow us to forecast and the ability of genetics to allow us to forecast, you're a sceptic. what do you make of the rise of what has been called a breed of super forecasters, and interestingly, the chief adviser to the current prime minister in the uk, borisjohnson, that is dominic cummings, he advised all of us journalists recently to stop chuntering and listen to the super forecasters. he referred to one particular individual in the united states, philip tetlock out in pennsylvania, a political scientist and psychologist, the founderfather of this notion of super forecasting. do you believe in super forecasting?

i interviewed philip tetlok for my book because i'm very captivated by his notion, and i agreed broadly with what he said, but i think dominic cummings may be missing a few details here in that when tetlock describes super forecasters, they are people with very open minds, they're not ideological and are constantly reviewing it, revising their opinions. i don't think anybody genuinely believes that that describes dominic cummings. you think he is harnessing super forecasting for a much more ideological project? yep and i think the heart and soul of super forecasting, and tetlock writes about this brilliantly, is that these are pretty ordinary people so they not kind of genetically supreme, shall we say, who read very widely, keep a very open mind, are super curious about the world,

notice all kinds of small things and, you know, don't come to a conclusion overnight. and they are willing to recognise when they get it wrong. isn't one of the key problems with your notion that we need to respect and even embrace some certainty, that it encourages endless argument and division about what is happening in the world today and more than anything else, i'm thinking of climate change. it seems to be one of the most important ways the public mind has changed on climate change as an issue is because of forecasts. forecasts of trajectory, where we are going in terms of the planet's dangerous warming but your advice to us all is, actually, don't get hung up with forecasts because they may well be wrong, don't take them seriously. i think there is a really important distinction to be made here which is we can know some things, and i would include climate

change on this, and epidemics, are generally certain. we know that epidemics will continue to happen, we know that climate change is real, let mejust finish if i may... everything else you've just said, how can we take seriously the fact you decided you know climate change forecasts are accurate. you've questioned so much else. so we know the preponderance of evidence says that climate change is real. we know, for example, the bank of england says there will be further crashes. we know that epidemics will continue to happen in human existence. that's generally certain but, and it's an important but, it's specifically ambiguous, which is to say we can't forecast which forests are going to catch fire next year, which houses are going to be flooded next year or even this summer and we can't say why or when the next banking crush will come — we know it will happen and therefore, instead of hoping to plan, assuming that these things won't exist, we do much better to prepare on the assumption that they will exist.

right. it's just you, margaret heffernan, write this in the book uncharted, "the hard truth is so much in life always will be unpredictable. we are better off accepting that than falling for propaganda." now, the climate change deniers describe the forecasts of doom for the planet unless we take dramatic action on reducing emissions, they describe that as propaganda. it seems to me you're making an argument which climate change deniers — just for one group — could well take hold of and advantage of. i think actually, the problem is different from the one you've described. i think because we can't predict exactly as it were, month by month, how climate change will play out, we've been stalled and done nothing for 30 years. and i think if we had been willing to accept this

is generally true, specifically ambiguous, but it represents such a huge threat that we can't afford to wait for the perfect data, if we had done that, we would be far more advanced in our fight against climate change than we are today and i think this obsession that we have to have the perfect plan before we can walk outside the door, i think this holds us back from doing the things we need to do and the things we could do and a perfect example of this is the coalition for epidemic preparedness, which although governments weren't terribly keen years ago, started up saying actually, because we know these things are going to happen, we need to start now putting in place the things that will need when they do. but again, it's problematic because you develop this idea of preparedness and being imaginative and creative about the different range of possible futures one needs to consider. right, which speaks to probability.

of course, i get all that and on one level it's commonsense but on another level, if you are in government, you can't afford to be investing in the preparatory infrastructure for a whole range of options because by definition some of those options will never come to pass on all the infrastructure you've invested in and the special measures you've taken will have been a waste of money and utterly redundant. i'm not saying you invest in everything obviously, that would be nonsense, but for example, after the banking crisis, what did we do? we make the banks keep more capital so when they were hit by shock, that we could not forecast, when they were hit by shocks they would be more stable. in other words we did exactly what i'm talking about which is, we said we don't know when the next crash will come, you don't know what will cause it, let's invest capital and making sure these institutions are more stable and one reason why today's stock market is going mad and will do for weeks to come, one reason the banks are stable is because of what we did then and all i'm saying is, you can take the same attitude

to other things when you know they are generally certain and specifically ambiguous and be better prepared than some people, places too. one extraordinary influential sector of 21st—century life and culture that you seem to have a downer on, see as a malign influence, is big tech. yes. you seem to be saying the more we subcontract our thinking, our analysis, our future projection and prediction of what the world is going to become to machines, the more we are rendering ourselves incapable of independent thought. and you seem to be saying big tech is encouraging us to allow them to make the key decisions in the future because it serves their self—interest. you see them as very malign. yes, which is ironic because i ran tech companies for years and i love a lot of technology but i don't really

love what it's become. and i think actually what's happened in silicon valley, the whole business model based on big data, if we had all the data in the world about you, we can predict what you want and that's how we make money. they can make life better for you, is their position. well, we can definitely sell you more stuff, let's put it that way. as we know everything about you, we think we can predict what you want to make money out of you by doing that. and actually what's happened is, it doesn't work, it doesn't work the way everyone hoped and as a consequence, as a consequence, you are seeing this shift to, ok, let's use all sorts of conditioning tools, some kind of carrots and some sticks to make you do what we want you to do because actually, we can't predict it. is it really, though, robbing you of agency because you're suggesting it is but at the same time i got a smartphone in my pocket which is a window to all of the world's knowledge. it gives me, it empowers me in a sense

to get away from traditional hierarchies, it gives me more of a voice in the world, notjust me but billions of other people. isn't that something to be celebrated ? certainly the tech ceos who you seem to have fallen outwith, think so. well. one of them, john booney, says tech is overwhelmingly positive force in the world for one reason, anyone with an internet connection has all the world's knowledge at theirfingertips, digital technology truly gives us freedom. well, i would absolutely argue with that. for example, what we can already see is that it's really damaging our power of concentration. we can see that the more you use gps, the less spatial awareness you have. we can see that we are becoming so used to distraction that we can't actually sit down and read books for more than five minutes before we have tojump up and do something. the truth of the matter is, all new inventions have a positive side and a negative side.

and we have a huge deluge of propaganda coming at us from silicon valley while at the same time realising that our ability, for example, to have conversations just like this, face—to—face, not through a machine, that is profoundly limited. i know kids today who can't actually communicate with each other face—to—face. they can't do interviews for universities, forjobs because they actually have never practiced it because they never needed to. i don't think that is a trivial loss. what you are saying overall seems to be a very important about human agency and whether we are actually gaining or losing as we progress through the 21st century. i just want to end with the thought that goes back to your earlier books about wilful blindness, when you talk about the way organisations, groups and communities, even individuals, they see bad stuff happening but often

they stay quiet about it. and you refer to everything from bankers during the financial crisis of 07—08, people involved in the occupation of the iraq. and i think you alluded to feminism and some of the things women put up within their individual lives. since you wrote the book, we've had metoo. right. does that give you hope in this sense, human beings are beginning to confront wilful blindness? yes. and i also think human beings have a phenomenal capacity to experiment, to invent, to explore and in many ways what this book is saying is, don't rely on forecasts to tell you what to do. think for yourself, look around for yourself. have conversations, take risks. invent different things, explore by yourself because actually human agency is what has created all the best things in our world today. and we need to cherish it and protect it and not let it be taken from us by people who purport to know

exactly what is going to happen tomorrow. a very interesting note to end on. margaret heffernan, thank you so much for being on the hardtalk. my pleasure. thank you. thanks a lot. hello there. for this upcoming week, it looks like much of the country will be fine and settled with some sunshine, thanks to high pressure over the baltic states. but we will have this weather front affecting the north—west corner of the country, that's going to bring strong winds and some persistent, at times heavy rain to the north—west of scotland. eventually weakening and sinking southwards across the country on thursday and then introducing colder, northerly winds to all areas

by the end of the week. but for much of this week it will be dry thanks to that area of high pressure, certainly across england and wales where we will have that rain in the north—west. the night will continue to be chilly as well. now, as we start monday it is going to be a cold one thanks to clear skies and light winds. a widespread frost away from the north—west where we will see more of a breeze and some cloud. but that frost could be quite hard in a few places and we could even see a little bit of mist and fog, too, as the winds will be lighter. but it does mean it's a chilly start to this morning, but there should be plenty of sunshine, certainly for england and wales. for the north—west of scotland, this weather front bringing outbreaks of rain to the northern western isles and thicker clouds tending to push into scotland and northern ireland through the day. it will be windy here as well with local gales, a breezier day further south although you have all that sunshine, which will tend to be a little bit hazy at times — it shouldn't feel too bad, with temperatures reaching 12 or 13 degrees, but always cooler around some southern

and eastern coasts. as we had the monday night, it stays cloudy, windy, outbreaks of rain across this north—west corner, really piling up across the outer hebrides and the north—west highlands. but here it won't be a cold night like it will be for south england and wales. it's a bit of a repeat performance of tuesday, england and wales largely dry, with some sunshine, albiet a little bit hazy at times. more cloud generally for scotland and northern ireland and it stays windy with rain really starting to pile up across the northern western isles in particular. we'll also import some slightly milder air up from the south and could see temperatures reaching 14—16 degrees in england and wales. the pressure pattern then for wednesday, little change, high pressure keeps things largely fine and settled for england and wales, our weather front bringing generally bringing more cloud to scotland and northern ireland, will start to weaken as it sinks south—eastwards, so there will be some spots of rain on it. further south, again, quite mild, those temperatures 11—14 degrees. but the milder air doesn't last. it gets squeezed out as we see a new area of high pressure build over the atlantic, that bring northerly winds right across our shores down from the arctic, so it will be settling down towards the end of the week but turning colder for all.

there will be some sunshine and also wintry showers across the north of the uk on that northerly wind.

this is bbc world news, i'm sally bundock. our top stories: coronavirus deaths soar across europe prompting leaders to bring in even tougher restrictions. stay apart or face tougher measures — britain's warned a lockdown is looming if social distancing advice isn't taken seriously. can plans for tokyo 2020 stay alight? sparked by international pressure, organisers consider a delay to the summer olympics. and delhi is left deserted as india tries to take on the disease — millions begin life under lockdown in the capital.