Artificial Intelligence is never far away from the news in 2017, with most reports highlighting the dangers of AI. There’s no shortage of celebrity scaremongers, including Tesla and SpaceX founder Elon Musk. “If you’re not concerned about AI safety, you should be,” he said last week. He added that AI presents “vastly more risk than North Korea.”
Well, not in the short term. AI is still decades away from attaining human-level intelligence – if it ever does. So right now there is much more risk in Trump versus Kim Jong-un.
Nevertheless, AI is a genuine threat to our jobs. And that’s regardless of whether you’re blue collar or white collar. The University of Oxford polled 352 AI academics and industry experts from around the world about when AI will outperform humans in a variety of jobs. The responses are indeed scary.
The experts predicted that AI will be able to drive a truck by 2027, work in retail by 2031, and work as a surgeon by 2053. All jobs will be fully automated in the next 120 years, the report concluded.
Whenever there’s a report like this about AI taking our jobs, techno-utopians invariably respond that humanity will invent new jobs. They point out that until a few years ago, there was no such job as Uber Driver. Go back a couple of decades and there was no such thing as a Web Designer. We’ll be okay, these utopians assure us: technology will provide new jobs as fast as it takes away the old ones.
Except that won’t be the case, this time. Take Uber drivers, for example. Once Uber puts a fleet of automated cars onto the road, many of its drivers will become unemployed. Some may transition into fleet managers, or re-skill and become Web Designers. But many, perhaps most, Uber drivers will find themselves out of work once driverless cars become commonplace (which is likely to happen within a decade).
If it was just Uber being disrupted, society would find a way to re-employ their drivers. The problem is, AI will disrupt a great many industries. There simply won’t be enough new jobs to go around.
Not everyone agrees that all professions are under threat. Local accounting software firm Xero claims that AI “won’t eliminate accountants, but empower them.” It says that AI will remove the tedious manual work, freeing accountants to “think and analyse” more. A Wired.com article states pretty much the same thing, but with a less optimistic conclusion. It claims that AI will soon automate tax reporting for most people, resulting in less accounting jobs.
What both Xero and Wired agree on is that AI will automate relatively simple accounting work. However, I’m inclined to agree with Wired that this will result in significant job losses. The jobs that remain will be more interesting, for sure, but that will be cold comfort to the tens of thousands of people who are made unemployed.
What can we do then to alleviate the concern about AI taking our jobs? This is where governments will need to step up. At the very least, we’ll be looking to government to provide a safety net in the form of income protection. There’s also the small matter of dealing with a new wave of economic inequality. Why? Because the only people guaranteed to gain money from AI are the people who own the companies that produce AI. Uber’s shareholders, for instance.
The most common safety net proposal is a Universal Basic Income (UBI) scheme. Here in New Zealand, we’re already familiar with government income support – we colloquially call it “the dole.” But in some places, notably the United States, a UBI is viewed as something akin to communism.
It’s hard to see how UBI can be avoided though. As former New Zealand Prime Minister Helen Clark put it recently, “guaranteed minimum incomes” will be needed.
Just as importantly, Clark also said that governments “must invest in capacities and opportunities for citizens.” Partly that’s to prepare us to work in partnership with AI, rather than compete with it.
In a Future of Jobs Report published last year, the World Economic Forum discussed some of the skills that will be needed in an AI era. Creative and interpersonal skills are high on that list. In the healthcare industry, for example, AI will automate things like diagnosis and treatment plans. But we’ll still need to translate and communicate that data to patients, preferably using the empathy that computers lack. So medical professionals will still be needed, but their required skillset will evolve.
Since AI will force many of us to re-skill in this manner, one of the report’s best recommendations was for government to find a way to incentivize lifelong learning. It pointed to Denmark, which allocates funding for two weeks of certified skills training per year for adults. The result has been a “very high degree of employment mobility” in that country. That’s something New Zealand could easily replicate.
Is AI as risky to society as nuclear weapons, as Elon Musk is warning us? It’s too early to tell. However what we can do now is prepare ourselves for when AI takes our jobs. That’s the real danger, at least for the foreseeable future.