Machines are poised to take over even top decision-making jobs as the artificial intelligence revolution gathers pace, writes Michael Odell
When the Luddites smashed the Nottinghamshire looms in the early-19th century they did so out of concern for the future of human labour. The Industrial Revolution was wreaking huge social change. But the economy, they soon learnt, would always win in a struggle between humans and efficient machines.
These days we accept that low-skill jobs are perennially eaten away by new technology. But what if a revolutionary order of new machines threatened the jobs of the decision-makers, the white collar “knowledge workers”? What if the machines came for you?
If you have used Siri on an iPhone, browsed the algorhythmically-derived recommendations on Amazon or even just simply checked in for a flight online, you probably know there has been an artificial intelligence or AI revolution underway for some time.
Computers can do simple bulk clerical tasks well. They have saved us time or money and the effects have been benign. And so what if a few jobs are being lost to the machines? Call-centre workers, travel agents, Amazon delivery drivers (soon to be replaced by drones or driverless vehicles), these are the low-skill loom workers of yesteryear, right? Anyone who complains is surely simply a Luddite 2.0.
Except that the robots are now ready for your job. They are no longer prepared to help out answering the phones or checking personal identification at reception. They want a seat in the boardroom, a stool in the lab and they don’t require a pay packet, holidays or pensions to do it.
Because of advances in two of the crucial areas of computing – natural language learning (computers are now able to derive meaning from human language) and machine learning (they are able to learn from data rather than merely respond to human programming) – a new generation of AI is challenging for the top jobs in law, pharmaceuticals, medicine, even government.
DIGITAL ASSOCIATE LAWYERS
In the law, AI has arrived in the form of the “digital associate”. Law firms are increasingly turning to predictive coding, software which uses powerful algorhythms to search electronic information for relevant legal documents. The legal task of discovery, the collating of relevant documentation for a case, is now often delegated to machines. So whereas even ten years ago a law firm might have employed hundreds of highly paid lawyers to search millions of documents in preparation for a large merger case, firms such as Blackstone Discovery, a Silicon Valley-based e-discovery firm will analyse, say, 1.5 million legal documents in a matter of days for less than $100,000.
There are other firms that go further and have developed AI which is able to collect potential evidence from an array of digital sources. For example, does the language used in an e-mail exchange or even a sudden termination of that exchange – perhaps the correspondents decided e-mail was too dangerous and continued their discussion on phones instead – suggest that illegality was being discussed? A firm such as San Francisco-based Cataphora has developed AI to monitor and analyse personal and organisational patterns of behaviour online.
“I just don’t think the law is keeping up with technology,” according to Skype’s assistant general counsel Jason T. Anderson. “Everybody’s got an iPad at home, a computer at home, and a computer at work and a mobile phone. How to keep up with all this chat and collect data from these locations is a huge problem.”
Perhaps unsurprisingly AI has made fast inroads in the world of finance too. Large-scale stock trading has been undertaken by robots for over a decade. A software glitch caused the notorious Wall Street “flash crash” of 2010 which wiped 10 per cent of the Dow Jones industrial average in a few minutes.
VIRTUAL MARKET ASSISTANTS
However, the latest developments are not aimed at enhancing the speed of trading, but actual market prediction. American firm Kensho’s new computer or “virtual market assistant” nicknamed Buffett, in wry tribute to American super-investor Warren Buffett, can answer verbal questions about complex market situations and offer detailed advice.
You cannot stop the rise of the machines – even the creative industries are not immune
For example, how will a coup in South America, where wood is sourced, affect the stocks of a major home improvement chain? How will regime change in Cairo impact on oil-specific stocks? Buffett is essentially pre-loaded with the causal ebb and flow of entire financial market history, and has the power to crunch through this data and see patterns which would take even a skilled human mind years.
And if you need to sit down and wade through Buffet’s findings at leisure, you could always turn to “Quill” a robotic journalist able to write detailed financial reports in fluent, engaging English.
We have known computers can deal with vast numbers faster than humans since the first electronic calculators in the 1970s, so the fact that they are able to handle big data sets is no surprise. The difference now is that, having analysed terabytes of information, in legal documents or financial reports for example, the computers can think, infer and make suggestions which is the creative white-collar bit that we like to think we are best at.
Even the reassuring, white-coated doctor at your bedside may soon have a disconcerting battery-power light on their forehead. IBM’s Watson computer is currently being used at the New York Genome Center to help create personalised treatment plans for 25 patients with brain cancer. The weeks it would previously have taken human technicians to match DNA profiles with mutations and matching drugs now takes Watson minutes. According to Niven Narain, president at Boston–based Berg Pharmaceuticals, the estimated $1 billion cost of launching a new cancer treatment could be halved using AI.
You cannot stop the rise of the machines – even the creative industries are not immune. Newspapers already run articles written by robots. In March the Los Angeles Times managed to get a report on an earthquake online within three minutes thanks to a “robo-journalist”. Even government may succumb; in Germany, child benefit claimants are already required to interface with AI assessors.
But that doesn’t mean you have to wait for the Monday morning when you find a guy with a metallic chassis sitting in your swivel chair, pinging your executive toy. Both skills can be anticipated and, with stealth, usefully harnessed.
In law, for example, those trained in predictive coding who are happy to explain and introduce cost-saving e-discovery techniques to their clients, are going to thrive. Lawyers with a tech background are still needed to design a search and collate results for presentation to clients in what is sometimes called “digital curation”.
“You need to hire someone who is a specialist in the underlying technology,” says Paul Starrett, a corporate officer at UBIC, an e-discovery and digital forensics firm. “The law will increasingly mesh with computers.”
Indeed some lawyers are already straddling the two worlds so that future legislation is actually conceived and written in digital form. The secret across all white-collar jobs vulnerable to Al innovation is to skill up, understand the technology and identify the crossover areas where human intelligence is still needed.
Some 17,000 years ago the cognitive revolution in homo sapiens – the sudden intelligence leap that allowed us to co-operate with others and think abstractly – suddenly put us at the top of the food chain after millions of years battling it out with Neanderthals, sabre tooth tigers and mammoths. Cognitive computing might do the same for machines in the job market.
A recent Pew Research Center study asked 1,900 tech experts whether they expected AI to create or destroy net jobs by 2025. Opinion was divided 50/50. The optimists remembered that, between 1810 and 1910, the number of Americans engaged in agricultural labour fell from 90 per cent to 2 per cent. The urban revolution found other things for them to do. Surely the AI revolution will do the same?
Starker voices disagree. According to Jerry Kaplan, a Silicon Valley entrepreneur and visiting lecturer in AI at Stanford University: “People don’t understand it [AI]; they don’t get what it’s going to mean… I feel like one of the early guys warning about global warming.”