At 13:07 on April 23rd, 2013, the official Associated Press Twitter account tweeted the following news item:
“Breaking: Two Explosions in the White House and Barack Obama is injured.”
At the time of the tweet, AP’s account had around two million followers. The post was favourited, retweeted, and spread. At 13:13, AP confirmed the tweet was fake. Three minutes later, then-White House press spokesman, Jay Carney, confirmed that there had been no explosion and that the president was alive and well. There were no bombs, and no-one had been killed.
The target was never the president or the White House. The target was the financial markets. In less than a split-split-second, the state-of-the-art computers running state-of-the-art high-frequency trading algorithms picked out the keywords from a trusted source and went into overdrive, reacting to what they perceived as a confirmed terrorist attack that never actually happened. The cost to the market? $136bn (£108.5bn) in around three minutes.
No human being could orchestrate an attack of that magnitude.
The tweet sent from the compromised AP account was claimed by the Syrian Electronic Army - a hacker group supported by the al-Assad government. A human might have spotted the tweet and, upon realising the catastrophic financial implications, called an AP rep for confirmation. But the whole point of the algorithms involved in high-frequency trading is to detect, analyse and make decisions faster than a human heartbeat. April 23rd, 2013 is an isolated example of what can happen when computers make mistakes that humans wouldn’t. Which begs the question: would you trust your transactions, bank accounts, portfolios and other financial assets to a computer? Perhaps more unsettlingly, will you have a choice?
Read more: Trading places: the rise of the DIY hedge fund
“There are two types of algorithms used in trading today,” says Juergen Schmidhuber, an AI researcher of more than 20 years whom theNew York Timesin November 2016 called the father of AI. “There are simple programmes pre-wired to do certain things that the traders have identified, for instance. Little tricks that enable it to propose a price for a certain share. Then, of course there are little [systems] that are just going through a bunch of rules, and depending on the risk profile of the client they make certain decisions according to not very intelligent self-learning mechanisms.
The WIRED Money 2017 speaker continues: “On the other hand, you have systems that have been used since the 1990s that learn from experience to become better prediction machines. These use neural networks to predict behaviour, financial indicators and so on. Then the hope is you have a system that works better than those of your competitors and detects patterns that the others don’t see.”
Schmidhuber’s work in the field of artificial intelligence is centred on deep learning neural networks. The systems he and his team at the Dalle Molle Institute for Artificial Intelligence Research in Switzerland are building are problem-solvers: they explore, probe and test independent of researcher interference and slowly - and then exponentially - learn to complete tasks. If you have an Android phone, you have a working example of Schmidhuber’s work in your pocket - it’s one of his lab’s projects that partly powers Google Voice. When it comes to fintech, the idea is to create smarter and smarter AI that helps finance work better for financial sector workers, investors and anyone who simply wants to figure out the best way to pay their mortgage.
Smarter computers, algorithms and dedicated AI systems are a fintech dream. Faster decision-making and deeper learning (recognising, for example, predictors of financial turbulence) are obvious and huge boons to financial organisations. But if your investment in the world of high finance only extends as far as tapping your debit card on a reader in the supermarket, things are about to change for you, as well. The panicked phone call to the bank’s Lost and Stolen department may become a thing of the past - as will the questions from your spouse about why and how you apparently withdrew £300 from an ATM in Ukraine last night while you were sitting together rewatching Game of Thrones. And once again: that future is dictated by smarter computing, and forms of artificial intelligence that can tell you-at-home from you-in-Odessa. In fact, the technology protecting you from financial fraud might have nothing to do with traditional plastic: in the not-so-far future, you’ll carry your financial security around your neck, on your wrist, or not at all.
“In the history of finance there has always been fraud,” says WIRED Money 2017 speakerJoern Leogrande, executive vice president of mobile services at Wirecard. “This is something that no technology will [overcome]... That being said, I think that wearable technologies are much more secure because they are directly connected to your arm. We see wearables that are connected to your heart rate. The security pattern is based on your cardiac rhythm, which is something that is much more secure that a fingerprint. We will see these kinds of technologies evolving in the next few years, related to biometrics like facial recognition and voice recognition. This will lead to much better security, especially when it comes to wearable payments.”
Making the switch from cash and cards to more high-tech devices has some obvious benefits. If you lose your wallet and the £20 note inside, you know the chances of recovering either or both are something smaller than minuscule. Lose your card, and in two minutes you’re on the phone to the bank to cancel it and receive a new one, usually (and inconveniently) ‘in three-to-five working days’. But paying via something more sophisticated - a smartphone, a wearable - can give you a greater level of security.
“The security of a mobile phone is much higher than your debit card,” says Leogrande. “When you use Apple Pay, you can combine this with TouchID. That’s a biometric process you don’t have on a classical card. If you lose your phone, you can try to find it with an app. You can’t do stuff like this with your card.”
But whatever their promises, computers aren’t infallible. As smarter payment methods evolve - especially through devices that are multi-functional, like smartphones - inevitably the lines between security and privacy will cross. We may not care much when Google or Apple tracks our GPS signal on our phones, because we need our Maps app to find the nearest restaurant. But how would you feel about retailers having access to the same information? If the AI running the local shopping centre can see the that you beeline for the branch of Lush, maybe you start receiving vouchers for Bath Bombs. That’s not so bad. But what if you’re spending your whole visit stocking up on new baby clothes? How would you feel about an intelligent computer system knowing that you or your spouse were pregnant - possibly before they did?
“In most cases, brick and mortar retailers know nothing about the consumers or their clients,” says Leogrande. “People come in, pay with cash or card, and they’re not on a loyalty scheme, so the retailers don’t know anything about their behaviours. This is about to change, and this change will cost huge investment. But from a retailer’s perspective, you have to do something about this because what we see - and this is pretty obvious - is the kind of power and investment technology companies like Amazon are stepping into… There is tremendous change coming.”
That “tremendous change” is already here, of course - online. ‘Recommended for you’ is an algorithmic idea we’re used to, thanks to the likes of Amazon. And financial advice is following the same path of analysis and prediction.
“I think the biggest change is that people are going to receive financial help before they even know it,” says WIRED Money 2017 speaker Nick Hungerford, CEO of Nutmeg, an online investment management company. “It’s a combination of big data and artificial intelligence. We’re going to be able to be more intelligent about people’s spending habits, their health, their lifestyles. [We’re] going to get more effective at predicting what they’re going to need for different scenarios of spending and saving - when people are likely to get married, when you are likely to have a baby, etc. So in five years, we should be really good at giving people financial advice before they even realise they need it.”
Ultimately, then, the challenge for fintech is one of consumer trust. We - broadly - trust Amazon. Our money goes in, things arrive on the doorstep. We broadly trust Google and Apple: we give them access to anything and everything on our phones, laptops and tablets, and in return they make our lives less bothersome. Fintech aspires to the same promise, and services like Wirecard, Nutmeg, Schmidhuber’s lab’s work on AI that can outthink a human banker or trader by a thousand-fold, are examples of how it can deliver. The multi-billion dollar question is: how much do we trust AI with our money?
This article was originally published by WIRED UK
