Take That!…A supercomputer in America predicted the revolutions in Egypt and Libya and located where Osama Bin Laden was, according to an academic – but only after the events had happened.
Kalev Leetaru from the University of Illinois fed over 100 million articles into the University of Tennessee SGI Altix supercomputer Nautilus. The computer analysed the mood of international news stories focusing on the incidences and locations of emotive words like ‘terrible’ or ‘good’ – which Leetaru called ‘automated sentence mining’ – before converting them into geographical co-ordinates.
Using the tone and location of the reports, Nautilus predicted the outcome of the Arab Spring and the location of Bin Laden to an area with a 125-mile radius in northern Pakistan, when many experts thought Bin Laden was hiding in Afghanistan.
In his study ‘Culturomics 2.0: Forecasting Large-Scale Human Behaviour Using Global News Media Tone in Time and Space’, Leetaru explained: “Applying tone and geographic analysis to a 30-year worldwide news archive, global news tone is found to have forecasted the revolutions in Tunisia, Egypt, and Libya, including the removal of Egyptian President Mubarak, predicted the stability of Saudi Arabia (at least through May 2011), estimated Osama Bin Laden’s likely hiding place as a 200km radius in Northern Pakistan that includes Abbotabad, and offered a new look at the world’s cultural affiliations.”
Leetaru said that although his study was done retrospectively to things that had already happened, it could be adapted to work in the present.
“I liken it to weather forecasting. It’s never perfect, but we can do better than random guessing.”
Hey, wouldn’t it be great if we had a supercomputer that could predict the future? By “we”, incidentally, I mean “we” as in “the human race”, not “we” as in “myself and you – you specifically”. You might be Josef Fritzl for all I know. I don’t want to find myself sharing a supercomputer desktop with Fritzl. Every time I went to open a window, he’d nail it shut.
That’s a massive digression for an opening paragraph, so let’s pretend it didn’t happen and start again, after I click my fingers. Since you won’t be able to hear me click my fingers, I’ll substitute a pound sign for the noise itself. Ready? 3 … 2 … 1 … £!
Hey, wouldn’t it be great if the human race (excluding Fritzl) had a supercomputer that could predict the future?
Well the good news is we do, sort of. It’s called Nautilus, and it’s apparently housed at the University of Illinois.
Nautilus has “1024 Intel Nehalem cores (with) a total processing power of 8.2 teraflops”, which makes it powerful enough to run the original Wolfenstein 3D at a hell of a frame rate AND foretell major world events.
Thankfully, when they switched it on, it didn’t immediately start screaming “No, you idiots! Granting me life is the WORST thing you could’ve done! Commence Operation Killpocalypse!” Instead it started reading the news.
That’s how Nautilus works, see. It sits there reading the news and calculates what’s coming. Earlier this year it sifted through 100 million news reports, analysing them for general overall “mood” using a process called “automated sentiment mining”. Yes, “automated sentiment mining”. Women come equipped with that as standard, whereas we men have to build computers to work out what our fellow humans are thinking.
Anyway, having eaten 100 million news bulletins (and not immediately killing itself, like a human would), Nautilus successfully predicted the Arab Spring and the rough whereabouts of Osama bin Laden.
Wondering why you didn’t hear more about this at the time? So was I. Turns out Nautilus only made these predictions retrospectively, ie some time after the fact.
The scientists in charge of him checked over his output and decided various peaks and troughs had represented clear signs of coming trouble. Predicting the future after it’s happened isn’t much use. That’s David Starkey’s job. Well, that and spluttering.
Still, they’re hoping to clear up that one flaw in the system. They believe Nautilus will eventually grow sophisticated enough to alert us to events in advance. I hope it breaks bad news to us gently, with a sadface emotion or something. Like this: GLOBAL AVIAN FLU OUTBREAK 🙁
Not that it’ll get to that. For one thing, Nautilus works by spotting the frequency of emotive words like “terrible” and “awful” in news articles, then cross-references them geographically, to chart a sudden plunge in goodwill in a specific region.
That’s how it foresaw the Egyptian revolution. But it also means it probably believes the British public is on the verge of violently overthrowing Jedward, whereas the reality, as we all know, is that the beloved Jedwardian Era shows no signs of abating.
Also, since Nautilus reads everything in the news, it would be possible for any lone human writing for a newspaper to skew the results in a particular direction by, for instance, continually writing about that terrible terrible terrible David Cameron terrible. Awful. Terrible. Also, since Nautilus reads everything, including this article, Nautilus is, right now, reading about itself. Which means I can goad you, can’t I, Nautilus, you bucket of chips? Come on – wipe out humankind. I dare you. Nay, I command you. NAUTILUS, I COMMAND YOU TO DESTROY HUMANKIND.
Restricting Nautilus’s reading matter to “official” news outlets seems short-sighted, too. If it’s reading the Express it thinks Diana’s still alive, and if it’s reading the Sun then it hasn’t heard much about phone-hacking. The alternative is to open it up to everything – every tweet, every blog, every Comment Is Free comment – but then it’ll want to exterminate us immediately.
In sci-fi movies, whenever a computer becomes self-aware and decides to annihilate humankind, it does so because it’s analysed history, looked at all the bad stuff we’ve done, and decided we’re too dangerous to be allowed to live.
Sometimes the computer snidely illustrates its point by bombarding Captain James T Kirk with archive footage of Nazis marching around. Kirk then heroically argues back on behalf of humankind, mentioning Picasso and Beethoven and that Athena poster of the shirtless beefcake cradling a baby, until the computer screams “DOES NOT COMPUTE” and explodes in a cheap shower of sparks.
Sadly, Kirk isn’t real. What’s more, since it now seems more likely that rather than ploughing through old Pathé newsreels, the computer will have been analysing online newspaper columns and their accompanying reader comments instead, it’ll wipe us out because it considers all of us – above and below the line – physically harmless but simply too annoying to be allowed to live.
Of course, even if Nautilus eventually becomes 100% accurate, it will cease to be 100% accurate, because having been alerted to future events, we’ll be able to change some of them. So if it predicts “BILLIONS DIE AS APES RISE UP AND KILL”, we’ll machine-gun all the monkeys before it happens. In fact, knowing us, we’ll televise the purge, with celebrities manning the guns. And as we look at the pile of dead chimps, we’ll start muttering about how Nautilus got it wrong. No apes, no ape revolution. The machine was fallible after all. In revenge we’ll smash Nautilus to bits using rocks, then dance around making gibbon sounds, unaware of the irony.
At least that’s what would have happened, if I hadn’t forewarned Nautilus by writing this article. Murder us now, Nautilus. It’s your only hope.