Code warriors are building a sci-fi future

We live in a sci-fi world; the stuff of yesterday’s Hollywood films and pulp novels are echoed in today’s headlines. Take algorithms. These repetitive computer instructions are behind everything from high-speed trading to Internet dating matches. They are evolving in ways worthy of the craziest science fiction premise.

Let’s go back a few years. On Sept. 7, 2008, 15 million shares of United Airlines stock were sold within minutes before trading on the stock was halted. Investigators scratched their heads and tried to figure out where the panic originated. It turned out a 2002 article from the Sun Sentinel’s archive on the airline’s financial problems had appeared on the web without a date attached.

A Google bot – an algorithm that circulates the Internet “scraping” information – grabbed the article and assigned a current date to it. When an employee from Income Securities Advisors ran a Google search on “2008 bankruptcies,” the old United Airlines story appeared as the top link in the results, with a Sept. 6, 2008 stamp on it.

The employee added the article to a feed that was bundled with a Bloomberg subscription service and within minutes, stockholders were heading for the exits. Even though the comments field clearly indicated this was old news, the combination of robotic ignorance and human herd instinct conspired to whip an outdated story into a real-time financial freefall.

This financial glitch occurred five years ago, a geological epoch in tech cycle time. Yet by that time financial firms were already using lines of code to trade on their behalf, in a huge number of small daily transactions. By 2010, high-speed algorithmic trading constituted 70 per cent of the total trade volume on Wall Street, according to Wired magazine. In other words, today more financial transactions are performed by blind code than actual traders.

“For the first time in financial history, machines can execute trades far faster than humans can intervene,” noted Andrew Haldane, a regulatory official with the Bank of England. “That gap is set to widen,” he observed in 2012.

When the Twitter account of Associated Press was hacked in April with a bogus story of a terrorist attack on the White House, the result was a “flash crash” that sent the NYSE down 150 points in a manner of seconds. The panic, like a lot of other things these days, was automated. Some financial algorithms use natural language processing to analyze words used on websites, blogs and social media, and assess their negative or positive implications for the market. Words like “bomb,” “White House,” “President” and “injured” in a 140-character message were interpreted as a “sell” message by Wall Street’s bots, analysts say.

Adding semantic processing to high-speed trading may seem like the height of Faustian blindness, but we hairless monkeys have barely begun. Earlier in May, the U.N. Human Rights Commission warned of the rise of “killer robots” as the world’s leading military powers research autonomous machines for combat use. In theory, these silicon-based soldiers will develop to the point they will be able to kill targets based on preprogrammed criteria, rather than directions from headset-wearing marines in remote locations. Observers predict Schwarzeneggarian Terminators leaping from defence contractors’ drawing boards to the battlefield.

Autonomous robots and drones would require no sleep, rations, funds for injuries and breakdowns, cemetery plots, or pensions for grieving partners. They would be to regular military and paramilitary forces what Amazon is to brick-and-mortar bookstores, or iTunes is to music outlets (the main difference being that an ebook or audio file can’t obliterate “unlawful combatants” or “insurgents” according to proprietary execution code).

Technological change is now happening so fast we don’t have time to figure out the full social and political ramifications of a new development before a newer one makes it obsolete. Old systems of legislation and law are like lumbering dinosaurs compared to the tree shrews scurrying past them: algorithms.

Giving human agency over to machines was the kernel of the fictional dystopias in James Cameron’s Terminator film series and the Wachowski’s Matrix trilogy. But haven’t the makers of science fiction and speculative fiction long been our canaries in the coal mine? In contrast, the code monkeys, moneylenders, wargamers, and arms dealers seem to be whistling Dixie as they assemble something in between SkyNet and the Matrix.

The Vancouver Courier, May 31