Realm of the SCENSCI

If an analysis of computing texts were to be done, I am willing to bet that ‘Alice in Wonderland’ would rank at the top (or near the top) of the list of most highly quoted books by computer scientists (think Chapter beginnings or fancy quotations below titles etc, or any place where us computer guys want to show that we too are sophisticated and have read a bit of literature). As a result of the debate over “Fake News” and related issues, this obsession with “Alice in Wonderland” seems ominously prophetic. This is because many computer scientists (the entire field to some degree) now find themselves to be in the roles of both the bewildered Alice, and the dangerous Queen of Hearts, in the wonderland that is sociopolitical life. This is unfortunate because human affairs in general and our-right-to-diverse-news in particular, are matters too sensitive to be left either to the clueless Alice who understands very little about this wonderland, or to the ruthless Queen of Hearts, who would want to rule it with an iron-fist. So without further ado, first:

Alice the Clueless Computer Scientist

Before discussing how computer scientists would go about handling fake news, let me share an anecdote that might shed some light on this matter. While working at EPFL, I had to review a paper written for a prestigious Data Engineering conference. The paper dealt with using crowd-sourcing for assisting automatic information extraction for the construction of very large knowledge bases, with facts on millions of entities from text corpora and Web sources. In the section detailing the limitations of their work, the authors used the following excerpt from the famous movie, The Slow, the Fast, and the Clumsy* :

“With a single gunshot, Brunette severs the rope, dropping Chico face-first onto his share of the gold. Brunette smiles and rides off as Chico curses him in rage, shouting, ‘Hey Brunette! You know what you are? Just a dirty son of a bitch!’”

The authors then asked if the above should lead to a fact Brunette helped Chico or Brunette saved Chico or to the fact that Brunette likes Chico or perhaps instead to Chico likes Brunette ? This they concluded was a limitation of crowd-sourced systems of their kind since humans can differ in their opinions on these matters. They added that the solution is to produce more advanced fully automatic systems that do not depend on humans, and can determine the real fact of the matter!! Their (false) assumption of course was that there is some objective truth that can be reached on such matters by an objective machine. Or to use the jargon, they believed that there was some ‘ground truth’ that could be obtained (more on ‘ground truth’ below).

So how would a computer scientist determine what is Fake News and what news is Authentic? They would need to have something called ‘ground truth’ available, where ground truth, as the name implies refers to the truth on the ground. For instance: the prime minister of the UK lives at Ten Downing Street; there are twenty cars parked outside my house, etc. The general idea behind using ground truth would be that there is direct knowledge of what is right and what is false (i.e., what is the ground truth), and then your computer program tries to approximate this ground truth. In the case of news, this will be done to classify news into the categories of fake and authentic. The bad news for this scenario, however, was pointed out by Socrates two thousand years back: while things such as 2 + 2 might have clear answers, trying to answer questions such as what is good or beautiful is not easy. And it appears that Fake news, despite initially seeming like 2+2 = 4, surprisingly falls into the latter category. Don’t get me wrong: some fake news is simple enough to determine. If the Guardian blatantly lies about Julian Assange like this:

It can be easily verified that this news is fake by reading the original interview (the fake summary of which Guardian presented) or perhaps reading the interviewer’s tweets:

But we realize rather quickly, if we are not blind, that things get tricky when dealing with more “complicated” stuff. What if one were to say that the US invaded South Vietnam under John F Kennedy? Or consider a hypothetical news item that reads, “The US aggression in (not defense of) Vietnam has the country devastated even in the year 2017”. Is this fake news? How does one check the veracity of this news? One would not find any reference to a US invasion/attack/aggression in the mainstream media or even in the historical scholarship. Do we find a reference to such an invasion by reading Chomsky and other “martian”, “lunatic” thinkers? If we think the mainstream is right, then Chomsky surely is a martian lunatic, and a fake news machine to boot. If, however, we think that Chomsky is not a martian lunatic, then the abundant reference materials of mainstream scholarship (which should ideally have served as ground truth) carry as much weight as the social media fake news trolls that people hate so much.

Or take the recent screaming headlines in nearly all of western media: “Fidel Castro the Brutal Dictator dies”. The news “Fidel Castro, the dictator dies” can be easily verified. However the news as it was reported always had the words “Terrible Legacy”, “Brutal” (or other words to that effect) thrown in. This would suggest that not all dictators are “brutal” and not all of them have a “terrible legacy”. A look at how western media covers dictators shows that indeed they do believe this to be the case. So dictators are:

A) not labeled brutal and talked in terms ranging from neutral to positive, such as, this gem of a piece on my country’s brutal dictator Zia-ul-Haq. Zia was in any brutality calculus a million (a bazillion?) times worse than Castro, and who tortured and killed opposition, and flogged journalists, and spread religious extremism in the country. (On a personal note, thanks to him, I grew up in China as my father became a blacklisted dissident journalist who couldn’t find a job anywhere in Pakistan)

B) the dictators’ brutality is downplayed and their legacy is praised, such as this ‘wonderful’ piece on Augusto Pinochet:

C) they are effusively praised and showered with glory. Indeed, like this (from the New York Times):

Finally, how, one may ask, are non-dictatorial brutal mass murderers, discussed?

Like this:

and this:

Of course, these are but a few examples. The record is filled with mass murdering dictators and democrats being praised for decades. If all the above are not brutal, then isn’t “Castro the Brutal dictator dies”, an example of Fake news? Furthermore, what about his altruistic actions, which are both innumerable and momentous, especially for those of us in the third world? Do they count for anything? Do they or don’t they enter the calculus of brutality? How exactly would an algorithm determine what is true and what is fake? Long story short, it won’t. There is no ground truth that an algorithm can approximate/determine, even for these plain facts.

Thus, when those most clamoring for the eradication of fake news have no consistency on what fake news is, then what chance do poor computer scientists have. The fact is that there are no solutions to this problem save for the following unglamorous, boring and tedious solution: the Chomskyan painstaking effort of repeatedly, repeatedly, and repeatedly, repeating repetitiously that the US invaded South Vietnam under Kennedy among countless other facts that martian lunatics such as Chomsky keep re-repeating, like the argumentative and opinionated bores that they truly are.

So it should be clear that computer scientists cannot algorithmically stop fake news. But people, especially the Democrats and liberal elites these days are clamoring for putting an end to this “menace” and everyone seems to be up in arms against this “newly discovered” problem. And this new found zeal against fake news is precisely why computer scientists are in danger of becoming the Queen of Hearts, even more unfit to delve into such matters than the clueless Alice!

Since something has to be done, or so shout the liberal elites, then surely something will be done, despite all the concerns outlined above. We shall devise our algorithms! But will our algorithms trust mainstream media or martian lunatics Chomsky, I.F. Stone and co? More importantly, who will tell our algorithms which source to trust? We could give weights to various sources. But who is we? The power to shape these algorithms would ultimately rest with powerful tech companies led by billionaires, who hold sway over, and are in turn swayed by, other powerful elites. Do we really want that? Who will be most hurt by this? You can bet it won’t be giant media corporations who will keep on doing just fine. It will be the small media outlets, the independent bloggers and other such, who would be most adversely affected. Paraphrasing Lewis Carroll, for every news story that is deemed fake by an arbitrary standard applied by our tech giants, the proclamation shall be: ‘Off with it’s head! Its too fake’

We can already see this happening:

So in conclusion, liberal elites need to get some perspective on “Fake News”:

1) It has always been there and both the liberal and right-wing news media have gorged on generous portions from the Fake News cake, so this sudden obsession is unseemly at best, hypocritical at worst. Admittedly, a section of Trump supporters has disseminated wildly repugnant stories, but to the liberals I will say this: Just because Clinton lost, it does not mean we end up requesting Mark Zuckerberg, Page and Brin, and co., to assume even more power than they currently have and become the final arbiters of what’s fake and what real. Then we are truly doomed. The Internet and Social Media for all its flaws, finally give a voice to small publishers and independent people. Surely, we get ALL kinds of crazy, fake, repugnant stuff on there, but Mark Zuckerberg is not the solution. Ask the lefties who for decades have been dealing with ‘fake news’. The only solution is point (2) below

2) There is no way to magically stop fake news apart from boring, sober, repetitions of facts as we see them. That’s it! But don’t we all know that that amounts to preaching and is so-uncool? If so, then yes, Mark Zuckerberg is the only solution. But be warned: in wonderland, the chickens always come home to roost!!

–End