Search This Blog

Wednesday 17 April 2019

'Calling bullshit': the college class on how not to be duped by the news

Professors at the University of Washington say the course provides the most useful skill college can offer writes James McWilliams in The Guardian


 
‘Our world is saturated with bullshit,’ the professors say. ‘This is our attempt to fight back.’ Photograph: Leland Bobbe/Getty Images/Image Source


To prepare themselves for future success in the American workforce, today’s college students are increasingly choosing courses in business, biomedical science, engineering, computer science, and various health-related disciplines.

These classes are bound to help undergraduates capitalize on the “college payoff”, but chances are good that none of them comes with a promise of this magnitude: “We will be astonished if these skills [learned in this course] do not turn out to be the most useful and most broadly applicable of those that you acquire during the course of your college education.”

Sound like bullshit? If so, there’s no better way to detect it than to consider the class that makes the claim. Calling Bullshit: Data Reasoning in a Digital World, designed and co-taught by the University of Washington professors Jevin West and Carl Bergstrom, begins with a premise so obvious we barely lend it the attention it deserves: “Our world is saturated with bullshit.” And so, every week for 12 weeks, the professors expose “one specific facet of bullshit”, doing so in the explicit spirit of resistance. “This is,” they explain, “our attempt to fight back.”

The problem of bullshit transcends political bounds, the class teaches. The proliferation of bullshit, according to West and Bergstrom, is “not a matter of left- or rightwing ideology; both sides of the aisle have proven themselves facile at creating and spreading bullshit. Rather (and at the risk of grandiose language) adequate bullshit detection strikes us as essential to the survival of liberal democracy.” They make it a point to stress that they began to work on the syllabus for this class back in 2015 – it’s not, they clarify, “a swipe at the Trump administration”.


There has been considerable debate over what exactly qualifies as bullshit


Academia being what it is (a place where everything is contested), there has been considerable debate over what exactly qualifies as bullshit. Most of that debate centers on the question of intention. Is bullshit considered bullshit if the deception was unintentionally presented? West and Bergstrom think that it is. They write, “Whether or not that usage is appropriate, we feel that the verb phrase calling bullshit definitely applies to falsehoods irrespective of the intentions of the author or speaker.”

The reason for the class’s existence comes down to a simple and somewhat alarming reality: even the most educated and savvy consumer of information is easily misled in today’s complex information ecosystem. Calling Bullshit is not dedicated to teaching students that Fox News promotes “fake news” or that National Enquirer headlines are fallacious. Instead, the class operates under the assumption that the structures through which today’s endless information comes to the consumer – algorithms, data graphics, info analytics, peer-reviewed publications – are in many ways as full of bullshit as the fake news we easily recognize as bogus. One scientist that West and Bergstrom cite in their syllabus goes so far as to say that, due to the fact that journals are prone to only publish positive results, “most published scientific results are probably false”.




Why smart people are more likely to believe fake news


A case in point is a 2016 article called Automated Inferences on Criminality Using Face Images. In it, the authors present an algorithm that can supposedly teach a machine to determine criminality with 90% accuracy based solely on a person’s headshot. Their core assumption is that, unlike humans, a machine is relatively free of emotion and bias. West and Bergstrom call bullshit, sending students to explore the sample of photos used to represent criminals in the experiment: all them are of convictedcriminals. The professors claim that “it seems less plausible to us that facial features are associated with criminal tendencies than it is that they are correlated with juries’ decisions to convict”. Conclusion: the algorithm is more correlated with facial characteristics that make a person convictable than a set of criminal inclinations.

By teaching ways to find misinformation in the venues many of us consider pristine realms of expertise – peer-reviewed journals such as Nature, reports by the National Institutes of Health, TED Talks – West and Bergstrom highlight the ultimate paradox of the information age: more and more knowledge is making us less and less reasonable.


  ‘Even the most educated and savvy consumer of information is easily misled in today’s complex information ecosystem.’ Photograph: Ritchie B Tongo/EPA

As we gather more data for mathematical models to better analyze, for example, the shrinking gap between elite male and female runners, we remain as prone as ever to misusing that data to achieve erroneous results. West and Bergstrom cite a 2004 Nature article in which the authors use linear regression to trace the closing gap between men and women’s running times, concluding that women will outpace men in the year 2156. To take down this kind of bullshit, the professors introduce the idea of reductio ad absurdum, which in this case would make the year 2636 far more interesting than 2156, as it’s then that, if the Nature study is right, “times of less than zero will be recorded”.

West and Bergstrom first offered the class in January of 2017 with modest expectations. “We would have been happy if a couple of our colleagues and friends would have said: ‘Cool idea, we should pass that along,’” West says. But within months the course had made national – and then international – news. “We have never guessed that it would get this kind of a response.”

To say that a nerve has been touched would be an understatement. After posting their website online, West and Bergstrom were swamped with emails and media requests from all over the world. Glowing press reports of the class’s ambitions contributed to the growing sense that something seismic in higher education was under way.

The professors were especially pleased by the interest shown among other universities – and even high schools – in modeling a course after their syllabus. Soon the Knight Foundation provided $50,000 for West and Bergstrom to help high school kids, librarians, journalists, and the general public become competent bullshit detectors.

In 1945, when Harvard University defined for the nation the role of higher education with its report on General Education in a Free Society, it stressed as its main goal “the continuance of the liberal and humane tradition”. The assumption, which now seems quaint, was that knowledge, which came from information, was the basis of character development.

Calling Bullshit, which provides the tools for every American (the lectures and readings are all online) to disrupt the foundation of even the most trusted source of information, reveals how profoundly difficult endless information has made the task of achieving that humane tradition. How the necessary shift from conveying wisdom to debunking it will play out is anyone’s guess, but if West and Bergstrom get their way – and it seems that they are – it will mean calling a lot of bullshit before we get to the business of becoming better citizens.

No comments:

Post a Comment