Whether the discussion is focused on an AI actress seeking an agent, “grief bots” that are virtual replicas of deceased loved ones, or if artificial intelligence is a bubble about to burst and take the stock market down with it, talk about this technology is definitely part of our current zeitgeist. 

Much of the conversation, though, is speculation about how AI might disrupt the way we work. A recent Leger poll found that 78 percent of respondents believed artificial intelligence could lead to job losses for people in Canada, and, since we can already point to an entire job category (junior software developers) that doesn’t appear to exist in Canada anymore, these worries may be top of mind.  

It’s not only about job loss, either. There are also concerns about AI surveillance in the workplace and algorithmic scheduling, which could make work more chaotic, as well as the near certainty that many will have to learn AI skills to retain their jobs.  As such, it’s understandable that many of us are experiencing some level of “AI anxiety”—the latest manifestation of “technostress,” a term that has existed since the mid-’80s, when computers first became a regular feature of office life. 

Since technostress can be triggered by insecurity and uncertainty, as well as the added work associated with learning new systems at work, AI anxiety might be poised to make the technostress of the 1980s era look like a walk in the park.

The case of the “missing” coding jobs

“The potential for disruption is much greater than with previous technologies,” says Dr. Wendy Cukier, professor of Entrepreneurship and Innovation at Toronto Metropolitan University (TMU). “And we’re already seeing it decimating entry-level jobs in services like call centres and programming, which impacts young people most.”

For the past 20-ish years, the standard career advice has been “learn to code,” as software programming skills were in demand. That all changed with the introduction of generative AI (GenAI) in 2022. 

“ChatGPT got to a point where you could essentially provide snippets of code that weren’t working and ask AI to fix it for you,” explains Dr. Shion Guha, assistant professor at the University of Toronto in the Faculty of Information. “The next step was to get ChatGPT to produce the whole thing. So now some people prompt ChatGPT to generate their code for them.”

This is known as “vibe coding,” a development that has made it next to impossible for entry-level software developers to find work in Canada. But AI didn’t act alone. “Employers are not hiring them and are preferring applicants with more experience, because they’re betting heavily that entry-level programming jobs can be completed by AI,” says Guha. “However, whether that is true or false is yet unknown.”

Who’s next?

So, are coding jobs a canary in the coal mine? There’s no consensus yet, partly because there’s still uncertainty about how capable GenAI will become.

Guha says GenAI represents Big Tech’s plan for a new revenue stream, since enthusiasm for previous products, such as social media and cryptocurrency, is waning. 

“Big Tech essentially wants to sell these AI products to everyone,” he adds. “And one of the best ways of marketing to other businesses is to tell them you have a product that will replace workers.”

In some offices, though, GenAI is, instead, making extra work for humans, who have to sort through poor-quality AI-generated documents known as “work slop.” These error-ridden documents often contain misfires, some of which stem from “AI hallucinations” (it literally makes things up) and, according to a recent Harvard Business School study, are destroying productivity.

“There are a lot of junior positions at risk of being cut,” says Guha. “What I would not buy into are claims being used to sell AI products, such as being on the verge of Artificial General Intelligence (AGI, a theoretical idea that this tech will one day rival human intelligence and become sentient), because we’ve kind of hit the limits of how much training you can do, at least for now.”

Rage against the machine: Technostress and AI anxiety

Much of the media’s focus has been on job loss. The other part of the equation, though, is about workplace “augmentation”—using AI tools on the job.

Graham Dobbs, senior research associate, Innovation & Technology, at the Conference Board of Canada, says that some recent data suggest there might be more of the latter than expected. Fewer of the jobs the researchers expected to disappear actually did. At the same time, an increasing number of existing and new jobs are now demanding AI skills.  

“The people we thought were the most likely to be replaced were probably basically the domain experts on how to best use AI with their current tasks and responsibilities,” Dobbs explains. “So, I would say subject matter and domain knowledge in the role you are in is still critical. That said, knowing how to augment it with AI is really important.”

Learning to work with a new co-worker can be challenging, though. Especially one that isn’t human.

“We already have many technological steps for things like getting authentications and all the layers that organizations have added because of cybersecurity threats,” says Dr. Ariane Ollier-Malaterre, Canada Research Chair on Digital Regulation at Work and in Life at the Université du Québec à Montréal. “Everything is more complicated.”

And this “techno-complexity” leads to technostress, for obvious reasons. As the saying goes: “To err is human; to really foul things up, you need a computer.”

Ollier-Malaterre warns that, even when everything is working perfectly, GenAI might aggravate technostress for a range of reasons, including increasing the pace of work and surveillance “bossware”—spyware that keeps track of things like mouse movement and how many angry emojis you use in an email. And then there’s surge scheduling. 

“In retail, you could have very short notice before you have to go to work,” says Ollier-Malaterre. “Because they now have sensors that can see when there are more customers in the shop and will ask for staff to come in, in real time.”

Oh. Good.

Technostress is an additional layer of stress at work that has detrimental outcomes on people’s mental and physical health. Ollier-Malaterre says research shows it’s associated with poor sleep, emotional exhaustion, and workplace burnout.

The resistance will not be digitized

The first step may well be simply talking to trusted co-workers and supervisors. “Many people feel personally incompetent, but often, it’s a collective issue,” adds Ollier-Malaterre.

For those who work in unionized workplaces, she suggests talking to a union rep. In fact, talk to practically anyone you trust. Just make sure it’s a human with “real intelligence,” she says.

“Artificial intelligence can fake empathy quite well, but it cannot actually demonstrate compassion in the way that human beings need,” says Dr. Lauren Dwyer, assistant professor in the Information Design program at Calgary’s Mount Royal University.

ChatGPT is also supportive to a fault, she says, so it will usually validate all ideas, traits and ideas—good or bad.

“You need that human friction that forces you to face challenges and address them,” says Dwyer. “ChatGPT is there to help you avoid them. It may be free and accessible, but there’s a good chance it’s doing a lot of damage.”

The biggest number in the aforementioned Leger poll was about government regulation of AI—a whopping 85 percent of Canadian respondents are in favour of the feds stepping in.

“We need some type of regulation of AI,” says U of T’s Guha. “What that looks like, well, nobody has agreed upon, but, especially in the Canadian context, we need this, since we’re essentially dependent on American AI companies telling us what to do.” 

We have, for the first time, a Minister of Artificial Intelligence and Digital Innovation, a post held by Evan Solomon, a former journalist and MP. The federal government is making AI sovereignty a priority. Solomon has mentioned AI education, which many believe is crucial, since most people using AI at work now are self-taught and haven’t received any formal training. 

“For starters, we need basic literacy for every single Canadian,” says TMU’s Wendy Cukier. “And we have to be able to tailor the training to people’s needs and preferences according to demographics.”

Beyond basic literacy, some people should aim to learn AI skills for innovation and still others will want to learn “deep” AI development skills, she notes. There needs to be a wide range of training opportunities through communities, libraries, business associations and online certifications.

“I also think we need real leadership and real strategy on economic transformation, focusing on how to improve productivity and drive innovation and entrepreneurship,” says Cukier.

Small and medium enterprises are already a surprisingly important driver of the Canadian economy, she notes. Women-owned businesses account for one million jobs in Canada, more than three times as many jobs as are created by Canada’s big banks, combined.

We also shouldn’t discount expertise in other areas. It’s the combination of AI skills and domain mastery that’s giving certain employees the edge. Everyone we spoke with emphasized the importance of deep knowledge and stressed that the ability to conceptualize problems and communicate effectively was key to a lifetime of gainful employment.

“The leading experts in this field do not know what’s coming in the next five years,” says Lauren Dwyer. “Focus on the skills that you will be able to transfer to a variety of jobs.”

And don’t forget that there’s one skill AI can’t master: People skills.   

“The quote-unquote ‘soft skills’ that we end up developing, like communication, empathy, caring, compassion, time management, all those are going to matter significantly,” Dwyer adds.

“The technical stuff, we can figure out. Maybe AI will help us master the tech. Maybe it won’t. We don’t know. But who you are as a person matters more.”