Last July, over a thousand of the world’s brightest minds in computer science and robotics, including Stephen Hawking, Elon Musk and Steve Wozniak, convened in Buenos Aires for the International Joint Conference on Artificial Intelligence. Held once every few years since 1969, the event hosts discussions on the future of AI — where it’s going, what it means for humans, and how best to prepare for the future of technology.
Historically, the talks have been mostly educational, relevant only to individuals and companies who understand the complexities of AI and who speak the industry jargon. The 2015 conference, however, had a decidedly darker, more urgent tone.
In a letter signed by all of the event’s attendees, the industries and nations of the world were warned of a Terminator-style nightmare, should AI technology continue to advance unabated (keep in mind, computing power doubles every 18 months or so, according to Moore’s Law). This sort of sci-fi fantasy remains, for most people and their daily lives, hard to take seriously. But, there is one very real reason to be afraid of AI and robots: the potential for wide scale automation, across all industries, to eliminate the need for human labor. In a future where AI and robots can do everything we do — even better than we can — what are humans good for?
There is one very real reason to be afraid of AI and robots: the potential for wide scale automation, across all industries, to eliminate the need for human labor.
Alastair Bathgate, the CEO of Blue Prism, a UK-based developer of software robots that automate repetitive back-office tasks like data entry and analysis, believes AI is a force for good. “We’re not trying to take jobs away and take the economy apart,” Bathgate said. “What we’re trying to do is transform the outsourcing sector, and to give people an opportunity to do much more rewarding work.” For example, if a company needs a data technician, Bathgate believes it will be easier (and cheaper) to run his software, rather than go through the process of outsourcing work to countries like China, where an estimated 3.2 million jobs have been outsourced since 2001.
Bathgate acknowledges that automation has an impact on jobs — just not on the enormously detrimental scale outlined in some predictions. “We’re going to have a shortfall in labor in most first-world economies, and we need to address that in order to maintain production,” Bathgate said. Among the most widely cited studies on AI and its potential impact on jobs is a 2013 Oxford study, which predicts that in 20 years, 47 percent of existing US jobs will be made obsolete by AI and robots. Kevin Kelly, founding editor of Wired magazine, takes that estimate even further, saying that in 90 years, 70 percent of jobs will be automated.
Bill Gates has stated publicly his concerns about the potential impact of AI on society and work, and Elon Musk called AI “our greatest existential threat.” There’s clear agreement throughout the tech world that AI and robots, in one way or another, will indeed change the way we work forever — but where people disagree is whether or not humans will be able to adapt. “Why don’t we just continue to reinvent ourselves?” Bathgate asked. “I have faith in the human race.”
“Foxconn will save money and create some new jobs, but it’s hard to imagine the number of new employees will be close to the number they let go.”
Others, like James Barrat, aren’t so optimistic. The author of Our Final Invention: Artificial Intelligence and the End of the Human Era, Barrat believes new jobs won’t be created fast enough to prevent widespread job displacement. “What are America’s five million professional drivers going to do the year after their jobs are taken by self-driving taxis, trucks and delivery vans?” Barrat said. “Doctors and lawyers won’t be off the hook either. Past technological advances have hit only narrow sectors of the economy. Because they compete for both rote work and brain work, robotics and AI will hugely disrupt every sector.”
And what about the new jobs that robots and AI will help create? The Industrial Revolution didn’t eliminate all working-class jobs; the same will be true of an AI/robotics revolution, right?
Not exactly. Barrat pointed out that the Industrial Revolution happened on a much smaller scale than what we’re about to see with AI and robots. Back then, it was mostly repetitive, blue-collar work that took the brunt of the industrializing impact, and it all happened at a more manageable pace. Workers had the time and resources to reinvent themselves. With the advent of wide-scale automation through AI and robots, whose implications reach far beyond blue-collar work, people might be scrambling — indeed, some already are — to find work. Barrat uses Foxconn, the world’s largest electronics manufacturer (the company is famous for making the iPhone), as an example; the company recently purchased 60,000 robots to replace 60,000 factory workers. “About one in twenty of those workers might get retrained to manage an automated factory floor. Maybe two in twenty could get a job building or repairing robots. That leaves fifty-one-thousand workers looking for something else to do, right away,” Barrat said. “Foxconn will save money and create some new jobs, but it’s hard to imagine the number of new employees will be close to the number they let go.”
While the Foxconn example may cause alarm, many people have good reason to believe humans will adapt to the changes brought on by automation. “You’ll see a similar transition from people doing menial chores, robotic in nature already, and then being released to be able to do high-value, more richly rewarding things,” Bathgate countered.
Rick Collins, President of Enterprise of Next IT, a developer of web-based virtual assistants, echoed the belief that the workforce won’t dwindle or be pushed into lower-paying jobs due to automation. Instead, companies will become more productive and efficient than ever before, and humans will then be able to dedicate themselves to the areas of work only humans are good at: management, creativity and anything that requires human interaction. “There’s always been something that’s come along that’s just supposed to turn the world upside down,” Collins said. “Every one of those things — the car, the plane, the computer — in fact did make a huge, huge impact. But people are still adapting and continuing to find different ways to use technology in life and work.”
In the past decade, certain jobs have already begun transitioning from the realm of human labor into the realm of machines. Customer support hotlines are now mostly automated voices, able to listen and respond to commands like the headset-wearing humans that once operated them. Amazon’s warehouses are buzzing with knee-high robots that pick up, move and drop off packages all day long, and do so without ever needing a break (except for the occasional recharge).
It’s also worth noting that AI has a long, long way to go before it matches, or even surpasses, human intelligence. Such an event, called the “singularity,” a concept developed in 1965 by British mathematician I.J. Good and popularized by sci-fi novelist Vernor Vinge, has been assessed by teams of government researchers to determine how likely it is to happen and whether or not it poses an immediate threat. The researchers concluded that it does not pose a threat, and won’t for a very, very long time.
Computers are getting incredibly smart and robots are becoming more dextrous, but the fact remains that humans are still required to operate them. They are machines with buttons and touchscreens and power cords — not artificially intelligent beings, capable of acting independently and producing original thoughts. They need us. And even then, as they rely on our own inputs and commands, computers and robots can make huge mistakes. Earlier this year, Microsoft set an AI bot loose on the internet, taking the form of a Twitter account named Tay. Its purpose was to converse with Twitter users by analyzing and learning from their speech. Quickly, Tay started slinging racist and sexist remarks, and even quoted Hitler. Microsoft was forced to take Tay down.
The point: AI and robots are getting better every year, and will continue to do so for the foreseeable future. But they’re still mostly reliant upon humans to function properly. And whenever they’re given the chance to work on their own, they can be pretty dumb. Should you be worried about a robot or a super-intelligent AI computer taking your job? Not yet. In 50 years? Maybe, maybe not — no one is conclusive on the matter. But for now, rest easy. Watson won’t be handing in a résumé anytime soon.