Technology has been a pandemic savior for many businesses. Automation, in particular, is both taking care of work that humans can’t physically do right now—for example, product picking in busy warehouses where social distancing is difficult—and handling routine tasks so that limited human resources can focus on more advanced priorities. In fact, if we put the right management policies in place, automation could help Canada’s economy become stronger and more productive post-pandemic, according to a new study from Joel Blit, an economics professor at the University of Waterloo.
There’s a very real juxtaposition of interests in that virtual reality.
Some of the industries most likely to be impacted by automation, robotics, and/or machine learning are the same ones struggling the most during the pandemic. People are uneasy about what technology means for their employment prospects. Research from PricewaterhouseCoopers found that in 2017, 37 per cent of workers worried about being replaced by technological advancements like automation. An International Federation of Robotics survey, however, found that a majority of respondents believed automation might help them get more highly-skilled jobs.
The impact of tech like automation and artificial intelligence will be unevenly distributed among the different sectors of the economy. New research from Statistics Canada shows that in 2016, 10.6 per cent of Canadian workers were at high risk of automation-related job transformation; 29.6 per cent were at moderate risk. Those findings are disproportionate for certain sectors: more than a third of office support workers and a third of Indigenous workers were at high risk—well above the percentage of people in the overall workforce.
According to research from the Brookings Institute, workers in jobs focused on physical and cognitive tasks that are routine and predictable are most vulnerable to being replaced by technology—for example, those in office administration, food preparation, transportation, and production. Sectors with high automation potential (over 50 per cent), as identified by Brookings in 2019, include accommodations, food services, manufacturing, warehousing, transportation, agriculture, fishing, retail, mining, oil and gas. Overall, lower-wage jobs are more susceptible to automation, as are men, marginalized workers and younger workers.
Many of those categories overlap with the sectors and workers most affected by the pandemic-related economic downswing. Historically, this isn’t surprising. Recessions lead to both an increase in automation and an increase in anxiety about that automation. That means many of the workers most negatively affected right now are facing a double whammy of pandemic impact. First, they lost work because of COVID-19; now they might never see those jobs come back.
Other concerns go beyond future employability. Machine learning algorithms rely on large amounts of data. Consumer protections like the CCPA in California put restrictions on data use and collection but it’s hard to shove that genie back in the bottle. And even with adequate amounts of data, algorithms are only as unbiased as the scientists who design them. Examples of bias in machine learning include facial recognition systems more likely to return false positives for people of colour and more frequent disabling of the accounts of Black people on Instagram.
As new and developing technologies find their way into various workplaces, they will have both positive and negative effects. Given the former, excitement about the potential is understandable. So is worker anxiety, given the latter. It’s important to address both the pros and cons of emerging technologies because some of the solutions embraced to help businesses get through the COVID-19 pandemic are likely to stick around for years to come. Here are a few of them.
Customer support teams for a variety of businesses are dealing with increased workloads during the pandemic. Several companies have turned to chatbots in order to boost their online offerings, provide service as employees transition to online work, or make up for distancing measures that lower the number of employees who can work on site at a call centre.
Just a few years ago, attempts to introduce chatbots quickly went south. But chatbot software has become more sophisticated in recent years, thanks to advances in natural language processing and their underlying artificial intelligence. For example, when the pandemic began in the U.S. PayPal dealt with more than half of its message-based customer inquiries via chatbots. And Mount Sinai, a health system in New York City, used an API-based platform (application programming interface) that enabled patients to chat with clinicians.
Robots have become a solution for some companies that can’t safely continue their normal operations during the COVID-19 pandemic. American company AMP Robotics told the New York Times that demand for its AI robotic recycling robots is up significantly. Robots are already being used for warehouse tasks like moving boxes but automating the more complex work of finding the correct items out of thousands is still more wish than reality.
It seems fair to assume that once the investments in this technology have been made, they will continue to be used into the future. The market for articulated robots is expected to grow significantly over the next five years: from $44.6 billion USD this year to $73 billion USD by 2025. The negative impact of the pandemic, which created issues with supply chains and production, is expected to be minimal.
Automated Content Moderation
Any business with a digital presence that allows some kind of user commenting has had to deal with content moderation. Some comments on websites and social media posts are merely annoying, but others can contain discriminatory speech that can harm your business’ reputation or may even be illegal.
Automated content moderation isn’t new. An example is when a business Facebook page specifies that comments with profanity should be deleted. This automated moderation became more valuable during the pandemic, freeing up employees for other tasks. While these systems are advancing, they are far from perfect. Automated content moderation has the potential to distance workers from moderating content that can be psychologically harmful, but when it doesn’t work effectively, it can target content in ways that exacerbate inequalities.
Contactless payment options like Square and debit tap have been increasingly popular in Canada in recent years, but the pandemic appears to have led to a bump in their adoption by even the smallest-scale retailers. Sixty-two per cent of Canadians are using less cash during the pandemic and 42 per cent avoid transactions that can’t be contactless, a Payments Canada study found.
Major credit card providers raised their upper limit for tap payments to $250 from $100 in April, and a Moneris study found that 40 per cent of contactless transactions are now over $100. Companies like Lush Cosmetics are not allowing in-store cash transactions at all right now.
Robotic Process Automation
The buzz around artificial intelligence and machine learning can make it all sound fantastical, but a lot of the automation actually in use in workplaces right now is done via robotic process automation, or RPA. By automating repetitive tasks like schedules, invoices, forms and notifications, RPA software allows staff to focus on complex work. For example, an Irish hospital used software bots to lessen nurse workloads, which was essential during the pandemic but will continue afterward.
RPA makes use of machine learning, but right now the tasks handled by the software are routine. However, routine tasks still need to happen and RPA can increase productivity. The result is an increasingly valuable sector expected to grow considerably in the coming decades. According to an analysis by Grand View Research, the global RPA market size was $1.4 billion USD in 2019 and is expected to have a compound annual growth rate of more than 40 per cent from 2020 to 2027. •