Powerful algorithms are slowly superseding humans in different aspects of our lives – robots and powerful machines are beating us at our own game. Shouldnít we feel frustrated, at the very least?
Uber CEO Travis Kalanick believes that humanoid robots will soon be delivering pizzas and chatbots will be entertaining passengers in driverless cars.
Kalanick isn’t the only one who shares this vision. Ambitious giant tech firms are likewise actively creating a similar scenario, with no signs of stopping.
Just last December, Uber launched its artificial intelligence (A.I.) lab to improve autonomous driving. Microsoft and Apple are using A.I. to power digital assistants, Cortana and Siri. Google decisively acquired DeepMind, whose A.I. program AlphaGo went on to defeat renowned 9-dan professional player, Lee Sedol.
Powerful algorithms are slowly superseding humans in different aspects of our lives – robots and powerful machines are beating us at our game, including our beloved Go! Shouldn’t we feel frustrated, at the very least?
The Battle Over A.I.
Billionaire inventor Elon Musk has no qualms about saving humanity with his SpaceX projects, including transporting humans to Mars in case A.I. “goes rogue and turns on humanity”, he said in an earlier interview with Vanity Fair.
“I think they’re really improving at an accelerating rate, far faster than people realise. Mostly because in everyday life you don’t see robots walking around. Maybe your Roomba or something. But Roombas [an automatic robot vacuum] aren’t going to take over the world”, said Musk.
The battle between tech revolutionists – who trumpet optimal living and prosperity using robots – and nonconformists struggling to defend the ordinary way of life has never been more apparent in recent times. And to say that we been forced to take an absolutely stance between the two – to adapt or to abandon – is an understatement.
It All Starts With Losing a Job
It is inherently human to be actively employed and enjoying that sense of satisfaction and accomplishment that comes with the job. It is a major key to happy living.
The 2017 World Happiness Report found that people with a job evaluate the quality of their lives much more favourably than those who are unemployed. The loss of employment translates not only to a loss of income, but also a major loss to one’s emotional well-being.
The “Hollowing-out” of Middle-Skill Jobs
A modern repercussion of A.I. is the mass automation of jobs, and thus the mass replacement of human workers.
“Computers and their robot cousins have increasingly displaced workers in accomplishing explicit, codifiable tasks”, says David Autor, an economist at the Massachusetts Institute of Technology.
What jobs are considered easily codifiable? Are blue-collar, manual jobs that require low-level skills most prone to automation? Not necessarily.
Some experts say that even white-collar jobs like sales assistants and lawyers are under threat of A.I., while cleaners and janitors are still immune to replacement. As computer scientist, futurist, and serial entrepreneur Jerry Kaplan wrote in his book, Humans Need Not Apply, “Automation is blind to the colour of your collar.”
Whether a job necessitates automation depends largely on its repetitiveness or challenge.
Two types of jobs are safe from being automated – for now.
The first are those involving manual tasks, such as food preparation, grounds cleaning and maintenance.
The second requires high-level skills found in professional, technical, and managerial occupations that emphasise “inductive reasoning, communications ability and expert mastery” – jobs that Autor terms as “abstract tasks”.
Imagine a scale (x-axis) from 0 to 100 that grades occupational skill requirement.
Manual jobs requiring low-level skills sit furthest on the left, middle-skill jobs occupy the centre position and high-skill jobs take the furthest right. As increasingly repetitive tasks (in middle-skill jobs) such as simple bookkeeping or clerical work become automated, available jobs in the middle-skill range shrink.
Amazon CEO Jeff Bezos pilots a giant mechanical robot at the 2017 Machine Learning, Home Automation, Robotics and Space Exploration (MARS) Conference.
How would the graph look when plotted against a y-axis showing employment growth over time? The graph resembles a U-shape.
The result appears intuitive: the demand for middle-skill jobs is shrinking. In 1979, four major middle-skill jobs accounted for 60 percent of employment. This has slid to 49 percent in 2007, and 46 percent in 2012. Simultaneously, low-skill and high-skill jobs have increased on either end. This phenomenon has been coined “job polarisation” by economists Maarten Goos from Utrecht University and Alan Manning from the London School of Economics.
The fundamental question remains: Who are the winners and losers in this era of automation?
Good News for Some, Bad News for Most
With the “hollowing out” of middle-skill jobs, employment in both high- and low-skill jobs has been growing rapidly. But does that mean employees in the latter two job types are benefitting from such change?
Not necessarily so.
Autor recorded a negative wage growth in the low-skill percentiles between 1999 and 2007. He added that workers displaced from middle-skill jobs had to compete with new entrants, creating an oversupply in labour and a resulting dip in wages due to decreased labour demands.
The scenario is different for those at the other end of the occupational skill spectrum – at least in the short run. Managers and professionals who exhibited inductive reasoning and effective communication abilities benefitted from the cheaper advanced technology, which boosted their productivity and output – skyrocketing the demand for their services.
With high demand and low labour supply, high-skilled workers “benefit from information technology” as it “should raise earnings in occupations that make intensive use of abstract tasks and among workers who intensively supply them”, says Autor.
But the future for high-skilled workers may not be so rosy. Over time, job losses have also entered the high-skill territory. Zero employment growth was at the middle of the occupational skill spectrum during the 1980s, sitting at the 45th percentile. Two decades later, it has shifted to the higher end of the skill spectrum, at the 75th percentile.
While there could be many forces at play, Autor believes that “automation, information technology, and technological progress in general are encroaching upward in the task domain and beginning to substitute strongly for the work done by professional, technical, and managerial occupations”.
With the rapid development in technology, Autor cautions that skills at a premium today may become routine tomorrow.
A New Wave of Eugenics?
Back to the long debated question: should we or should we not worry about the development of A.I.?
From the current trend of employment in conjunction with A.I., the level of intelligence seems to be the major, if not sole, criterion in determining value. In other words, smarter machines and smarter men will survive longer in the game.
Ironically, it sounds like a new wave of eugenics, albeit a subtle form, which echoes the Nazis’ attempts to target lives they deemed ‘unworthy’ for mass destruction.
Pragmatists will say that it’s an inevitable trend, as companies seek to make a profit at the lowest cost. “If automation is more cost-effective, firms will replace human employees with machines.” “Hiring” automated machines might appear more profitable in the short term, but the countereffects of this practice can come back to bite companies in the long run.
Firms make profits and distribute part of it to employees in commensuration with their contribution; employees use part of their income to purchase goods, boosting the demand for firms’ production. It is a healthy and self-sustaining system.
But with more jobs being automated, rising unemployment and declining wages would partially erode the system by hurting consumer spending and confidence. As Martin Ford explains: “Major fixed costs such as housing (mortgage or rent), health insurance, debt, food and energy will not fall even as income does fall. This will leave average households with less and less to spend on discretionary items — and that likely means weak demand for any business producing a non-essential product or service.
“And, hey, that’s most of the economy. Those businesses, in turn will see increasing pressure to lay off workers or further automate.”
Some economists argue that this is really not such a bad idea. Known as “post-workists”, these economists actually welcome the end of labour. To tackle the problem of poverty among displaced workers, they propose a world where the unemployed receive a “universal basic income” – a scheme fuelled by heavily taxing the rich who own capital, or machines. In this way, people will be free from work commitments, and have more time to pursue hobbies and engage in meaningful activities.
That sounds like a nice solution, but the post-workists’ vision creates more problems than it solves. The taxation and redistribution mechanism would inevitably engender unfair feelings among people who are still working and contributing to the economy. A further question is, would people spend their increased leisure time meaningfully? According to a 2013 study, it’s unlikely. Professor Mark Aguiar at Princeton University and his co-authors found that “about half of the work hours lost by US workers in the recent recession were reallocated to leisure activity, with most of this accounted for by sleeping and television watching”.
Will A.I. Overtake Us? If So, When?
The job has attracted many new medical school graduates for its ìrelatively high pay and regular work hoursî; also, ìradiologists generally donít need to work weekends or handle emergenciesî, says Martin Ford, a futurist and author of Rise of the Robots.
However, there is no easy path toward becoming a radiologist. One has to go through about thirteen years of education and training beyond high school. ìYou need to attend college for four years, and then medical school for another four. That is followed by another five years of internship and residency, and often even more specialised training after that,î explains Martin.
Despite the heavy requirement on education, ìit is conceptually quite easy to envision this job being automatedî, notes Martin. While the current technology is still far from replacing radiologists, visual pattern recognition software could help actualise the automation of medical image interpretation in the distant future.
In fact, IBMís supercomputer Watson, which heavily utilises A.I. techniques, is now being trained to ìreadî 30 billion medical images in the databases bought by IBM. The objective, though, is to assist doctors in diagnosing diseases faster instead of replacing them.
The meteoric rise of machine learning capabilities has allowed computers to continuously improve themselves – as long as they are given access to data.
“The A.I. doesn’t have to take over the whole Internet. It doesn’t need drones. It’s not dangerous because it has guns. It’s dangerous because it’s smarter than us,” said A.I. theorist, Eliezer Yudkowsky in an interview with Vanity Fair.
Elon Musk put forward the prospect that, with self-improving algorithms, even the most innocuous machines could develop into something less than benign.
“Let’s say you create a self-improving A.I. to pick strawberries, and it gets better and better at picking strawberries and picks more and more and it is self-improving, so all it really wants to do is pick strawberries,” said Musk. “So then it would have all the world be strawberry fields. Strawberry fields forever.”
When will computers be as smart as humans? Futurist Ray Kurzweil predicts it to be the year 2029.
In July 2015, over 1,000 leading researchers and experts in A.I. signed an open letter calling for a ban on offensive autonomous weapons, or colloquially known as “killer robots”. Among those who signed were Elon Musk, Stephen Hawking, Demis Hassabis and Apple’s co-founder Steve Wozniak. The question for the future is, how should the development of A.I. be regulated? Where do we draw the boundaries? Hopefully, we won’t have to rely on A.I. for the answer.
When will computers be as smart as humans? Futurist Ray Kurzweil predicts it to be the year 2029.
