Artificial intelligence (AI) will either destroy jobs or create new jobs — depending on which report you read. In truth, it’s probably a little of both — AI will undoubtedly replace human workers in some spheres, but it will also create new roles, many of which we can’t yet imagine. A recent report from PA Consulting, titled “People and machines: From hype to reality,” supports this theory and predicts AI is more likely to create jobs than destroy them.
The research behind this report, based on 750 cross-industry businesses in the U.K., found that 32% of respondents had invested in AI and automation in the last five years, split evenly across tools for cognitive and physical tasks. Of those that had invested, 43% reported an increase in jobs as a result, while 40% reported a reduction.
“Our research shows AI and automation are likely to lead to a net gain in job numbers,” the report reads. “As some types of jobs disappear, new ones will emerge.” These findings are supported elsewhere — the Organization for Economic Co-operation and Development (OECD) also predicts that while AI and automation will certainly impact jobs, there will be no net loss.
So what kinds of shiny new jobs can we expect? Sweden’s Einride gives us some idea — the autonomous trucking startup has announced that it’s now hiring remote truck operators. The company expects to make its first hire in Sweden in March, followed by similar hires in the U.S. later this year.
“New transportation system”
Einride, which has raised north of $30 million since its inception in 2016, has developed the electric T-pod, which it touts as an entirely “new transportation system.” Each pod is around 23 feet in length and can hold 15 standard pallets and travel 124 miles on a single charge. Most notably, the T-pod isn’t a standard truck retrofitted to drive itself — there’s no physical space inside the T-pod for a human to sit.
Einride’s T-pod adopts a hybrid driverless approach — on highways, the vehicle is designed to drive itself, but when it exits onto main city roads it switches to remote control and a human takes over from afar. Those same operators are also ready to control several pods on highways, should the situation require it.
Einride’s first hire will be a former truck driver who will retrain and learn to control multiple trucks from a remote room. The operator will also provide feedback on development of the company’s remote driver station, informing the work environment of tomorrow’s truckers. This evolution offers a glimpse into how other jobs could shift as AI advances — industry-specific expertise will likely still be required even if job descriptions change.
“Today, our autonomous pods are operated by developers — robot engineers trained to drive trucks,” Einride CEO and founder Robert Falck said. “A commercially scalable solution must rely on truck drivers, trained to remote-operate robots. The ins and outs of that future [are] what we’re investigating now, by involving truck drivers in the process.”
The global truck transportation market is pegged at $1.5 trillion, making it a sizeable target for disruption. Einride said electrification and automation will enable it to optimize the road freight industry by reducing fuel and energy costs by 70%, cutting operating costs by 60%, increasing productivity by 200%, and reducing Co2 emissions by 90%. Other companies working in the autonomous vehicle remote assistance and tele-operations space include Ottopia and Scotty Labs, which was acquired by DoorDash last year.
What all these companies show is that AI — for the foreseeable future, at least — will require humans to oversee things.
“AI will always have to be supervised,” Falck told VentureBeat. “Autonomous vehicles would be the obvious example, where I think human monitoring will be needed for many years to come, to help out in complex situations. And we’ll need rapid reaction teams to perform maintenance on broken down robo-cars. We’ll need all kinds of support functions that don’t exist today. But this is not just about practicality, it is about moral responsibility, which is a human construct and a notoriously tricky one at that.”
In a world full of AI, humans will likely still be required across industries, whether for supervision, remote assistance, training, ethical decision-making, or other input.
“Philosophers distinguish between explicit and tacit knowledge, tacit knowledge being difficult to transfer verbally,” Falck continued. “So how do you translate it into code? Machine learning could be said to be a solution to that problem, but that is not a transparent process. So we’ll need people who can reconstruct decisions made by self-learning machines, who can detect bias in self-learning algorithms, etc. The very opacity of AI and the ethical problems that flow from that will generate new interesting jobs.”