Depending on who you ask, AI and automation will either destroy jobs or create new ones. In reality, a greater push toward automation will probably both kill and create jobs — human workers will become redundant in certain spheres, sure, but many new roles will likely crop up. A report last year from PA Consulting, titled “People and machines: From hype to reality,” supports this assertion, predicting that AI and automation will lead to a net gain in job numbers. This is pretty much in line with findings from the Organization for Economic Co-operation and Development (OECD), a pan-governmental economic body spanning 36 member countries, which noted that “employment in total may continue to rise” even if automation disrupts specific industries.
Automation has gained increased attention amid the great social distancing experiment sparked by COVID-19. But it’s too early to say whether the pandemic will expedite automation across all industries. Recent LinkedIn data suggests AI hiring slowed during the crisis, but there are plenty of cases where automation could help people adhere to social distancing protocols — from robot baristas and cleaners to commercial drones.
Of course, any discussion about automation invariably raises the question of what it means for jobs.
Humans in the loop
As we’re still in the early stages of a broader shift to AI and automation, it’s not easy to fully envisage what new jobs could crop up — and which will be lost.
Slamcore is a London-based startup pushing to commercialize AI algorithms that help robots gain situational awareness from sensor data. Slamcore cofounder and CEO Owen Nicholson says we only have to look at some of today’s jobs to realize how difficult it can be to forecast the future.
“Contrary to some beliefs, I see robots as creating vast amounts of new jobs in the future,” he said. “Just like 50 years ago a website designer, vlogger, or database architect were not things, over the next 50 years we will see many new types of job emerge.”
Nicholson cites robot pilots as an example.
“Ubiquitous, truly autonomous robots are still a long way from reality, so with semi-autonomous capabilities with humans in the loop, we can achieve much better performance overall and generate a brand-new job sector,” he added.
There’s a growing consensus that humans will work in conjunction with robots, performing complementary roles that play to their respective strengths.
San Diego-based Brain Corp recently locked down $36 million to “help meet the growing demand for autonomous mobile robots (AMRs)” across industries affected by the pandemic — from health care to retail. Brain Corp is the company behind BrainOS, an operating system that integrates with hardware and sensors and serves as the “brains” for delivery robots used in warehouses, factories, and retail stores. BrainOS also powers self-driving floor cleaners that assist human workers. The machines come equipped with a range of sensors, including lidar and 3D time-of-flight (ToF) sensors, to self-navigate in dynamic environments.
Above: Robotic floor cleaners powered by BrainOS.
Brain Corp said demand for BrainOS-powered cleaning robots has surged amid the COVID-19 crisis, with retail usage growing 24% in April 2020 alone. “A significant percentage of this uptick — 68% — is occurring during daytime hours, showing that businesses are cleaning more frequently and operating the technology during peak times,” Brain Corp executive Michel Spruijt told VentureBeat.
The robots generate a significant amount of performance data, which is automatically compiled into reports that need to be interpreted, assessed, and analyzed to improve operation and fleet performance. While much of this work could be incorporated into existing roles, such tasks may eventually require dedicated employees, leading to the creation of new jobs.
“Managers can view the routes being cleaned, take a look at quantitative metrics such as run time and task frequency, and receive notifications around diagnostics and relevant software updates,” Spruijt said. “An understanding of these reports and how to successfully interpret and apply this data will be imperative in order to improve store operations using automated technologies.”
The robots are typically trained to follow routes through a “teach and repeat” method, with human workers guiding them along a cleaning route and making adjustments if the environment changes. As Spruijt is quick to point out, this process “proactively includes humans.”
“The robot is not a functional robot without the human,” he said.
Above: Cleaning equipment giant Tennant develops machines that use BrainOS.
Additional new jobs could include maintenance workers to ensure the AMRs are functioning properly.
“The process of physically building a robot and successfully maintaining it in the field requires a set of new or enhanced skills, which are likely to increase alongside adoption of AMRs,” Spruijt said. “As manufacturing lines start to ramp up robot production, [skill sets such as] tooling, light manufacturing, and familiarity with new hardware like touchscreens and lidars will be necessary. Once in the field, service providers with applied knowledge around robot maintenance and deployment are also important in ensuring success.”
Humans and robots have distinct strengths and weaknesses, which is why a human-in-the-loop model makes sense for companies embracing automation. Veo Robotics is a Waltham, Massachusetts-based startup that uses computer vision and 3D sensing to give industrial robots greater perception. Its Veo FreeMove system, which is due to launch next year, is designed to help manufacturers coordinate the best attributes of robots and humans, meaning it’s neither completely manual nor completely automated.
“What Veo does is enable a middle path, one where human workers with their flexibility, ingenuity, and dexterity can do the parts that humans are good at, while robots with their tirelessness and strength can help them by positioning parts or performing other elements of the process that are hard for the human worker,” Veo Robotics CEO and founder Patrick Sobalvarro explained. “It’s much quicker and cheaper to set up a work cell like this than to try to completely automate it because the human worker is able to do exactly the parts that are so hard to do automatically since they involve dexterity, sensing, and judgment.”
“This means skilled welders will spend more time welding and less time fixturing, and quality technicians will spend more time measuring and less time moving parts around,” Sobalvarro added. “Everyone will be more comfortable and get more products built.”
As industries strive to establish a new normal following the pandemic, human-robot collaboration could prove invaluable.
“Manufacturers have to reduce human density in factories to comply with these new [social distancing] rules,” Sobalvarro continued. “The human-robot collaboration that Veo’s system provides can address this, as it means that instead of having two humans working closely together in a work cell, you will have a human and a robot working together.”
Miso Robotics has been deploying its burger-flipping bots across the U.S. over the past few years. The company recently unveiled a next-gen robotic assistant called ROAR (robot on a rail) that can move between cooking stations, working the deep fryer and flipping burgers.
Miso Robotics ROAR
With widespread lockdowns, restaurants have been among the hardest hit by COVID-19. CEO Buck Jordan naturally believes automation will play a big role in helping the food industry get back on its feet.
“Now more than ever, [as] we are facing real new challenges and a whole new normal, navigating it is going to heavily fall on technology. And the restaurant industry is a prime example where automation needs to come in, in order to sustain the industry, drive growth, and create new job opportunities,” he said. “Incorporating automation into commercial food preparation empowers restaurant operators to safely reopen [and] attract customers with enhanced health and safety, as food comes into reduced contact with humans and points of contamination. [This] ultimately gives them the tools needed to increase production speeds and meet delivery and takeout demands — even in the face of new social distancing requirements that limit a full staff in the kitchen.”
This all sounds like a death knell for traditional kitchen jobs, but if restaurants aren’t able to meet safety guidelines, the reality could be much worse.
“Without giving restaurants the solutions they need to reopen and recover, the issue won’t be robots taking jobs. The real issue will be that there’s no jobs to take because restaurants can’t turn the profit they need to stay open, much less hire or create new job opportunities,” Jordan continued.
In the field
While drones incorporate various facets of automation, most require an operator to manage and oversee their deployment. People are needed to program flight paths and step in when things go wrong, and — as with other industries — perform maintenance.
The commercial drone industry was in ascendance before COVID-19 struck, with reports suggesting the market would grow more than fivefold by 2026 from $1.2 billion in 2018. However, the pandemic seems to have increased demand for drone services in areas such as medical supply deliveries and site inspections. Mike Winn, cofounder and CEO of drone data software platform DroneDeploy, told VentureBeat the company had expected demand to grow across more than 10 industries in 2020 — and COVID-19 has only heightened that interest.
“We’ve already seen growth in drone operations this year, despite COVID-19,” Winn said. “Our customer data showed 130% year-on-year growth among our active enterprise pilots, with minimal pandemic disruption from April 2019 to April 2020, and we just saw our greatest number of flights ever in May.”
The San Francisco-based company’s platform allows commercial drone operators to map, survey, and inspect aerial images. These can be used to harness data in industries ranging from agriculture to mining, construction, and insurance.
Above: DroneDeploy provides data capture and mapping software for commercial drone operators.
Even if automation doesn’t create many jobs off the bat, drones offer a glimpse into the way traditional roles may evolve, with field engineers gradually shifting into a completely new role.
“We’ve definitely seen companies training their teams on new skill sets involving drones,” Winn said. “For example, ‘drone operator’ is a quickly growing job title. Many field engineers are becoming drone operators as they have been wrapping drones into their roles more and more. In agriculture, this means capturing live field data, applying chemicals with precision, and more.”
Dimitri Onistsuk is the cofounder of Freedom Robotics, a San Francisco-based company that builds software to control and monitor fleets of robots. Onistsuk said he is seeing certain roles evolve, and companies may need to expand their workforce to include dedicated specialists.
“Companies need to broaden the skill range of their employees,” Onistsuk said. “For example, we’re seeing a bifurcation of robotics developers into engineers and operators. Because the demand for solutions is so high, it’s no longer a viable business decision to have engineers do operational tasks. Therefore it becomes important to have a higher-skilled individual focus on engineering and developing the robots — their vision, their algorithms, and so on — and have others focus on things like piloting, field operations, maintenance, and servicing.”
Companies may have to reevaluate their workflow to ensure people’s skills are being used effectively. “It’s not efficient to send your top engineers across the country to fix a broken robot that is just sitting on the floor of an impatient customer’s site,” Onistsuk added.
Like others in the robotics sphere, Onistsuk said Freedom Robotics has seen a “dramatic” increase in demand during the pandemic. More specifically, companies that were dabbling in the technology are now accelerating their plans.
“Customers have always taken the potential of robotics seriously, but now we’re also seeing a much greater sense of urgency, where they are no longer waiting years to get their prototypes just right and are instead rapidly moving toward deployment,” he said.
Onistsuk points to customers for real-world examples of how robots can augment the human workforce. These include teleoperator pilots and managers who supervise robots that are taking over some of the more mundane aspects of their role. “In the case of pilots, we are seeing humans in the loop, where a semi-autonomous delivery vehicle encounters an exception, such as a reflective material inside a warehouse or a bush that is difficult to navigate around near a sidewalk, and a human takes over via GPS, remote control teleoperation, or a script to reset back to autonomy,” he said. “In the case of manufacturing, robots are delivering parts to a particular station where their human counterparts take them and add them to the assembly.”
While it may be difficult to pin down the kinds of new roles that could emerge, it can be fun to speculate. Onistsuk considers a scenario in which a robotic snowplow struggles to tell the difference between a snowbank and a car underneath a layer of snow, requiring a remote operator to step in and interpret the scene and manually navigate if required. As new technologies mature in the coming years, we’ll begin to get a clearer picture of the ways AI and automation will impact the workforce.
“I think server and edge operations for robotics will be very important, where the necessary infrastructure is managed and assets are tracked,” Onistsuk said. “5G will bring a tidal wave of possibilities and will have a big effect on robotics and will require a new set of technical skills. I think there will also be a new generation of hybrid workers for jobs that are around today but that will be done collaboratively with a robot — sanitation, industrial inspection, surgery, and so on.”
A 2018 report from PricewaterhouseCoopers (PwC) noted that the industries likely to benefit most from AI are “human” and “highly technical” sectors, such as health care, education, and science. “Teaching requires high levels of interpersonal skills that cannot easily be replaced by AI systems or robots, although they can be complemented by them to meet projected rising demand for education over time,” the report found. “As such, we only expect around 5% of educators to be displaced by AI, more than offset by job creation of 10%.”
The report noted that machines could take on some of the more “boring” teaching tasks, such as marking homework or administering multiple choice tests. The report adds that the sectors “more likely” to experience net job losses are those with a “high degree of repetitive and routine tasks.”
One example is manufacturing, which PwC estimates will see 25% fewer jobs by 2037 as a direct result of automation. Meanwhile, 40% of existing “transportation and storage” jobs could be displaced by automation — due to a rise in driverless vehicles and automated warehouses — with less than half of those replaced by new types of jobs.
A recent report from the International Federation of Robotics (IFR) noted that there will be an “operational stock of almost 4 million industrial robots” in factories globally by 2022 and predicted “high demand for robotics skills” as part of the post-pandemic recovery process. The report added that governments will need to focus on education and training to equip their workforce with the necessary skills.
Even if AI and automation lead to net job creation across the board, there will be significant disruption and upheaval as the global workforce adapts to shifting demands. Some roles may become obsolete, while others may branch into new directions or lead to the creation of entirely new job titles. What is easier to predict, however, is the emphasis these changes will place on upskilling and retraining in the years ahead.