Most of us get the same basic questions in our employee performance reviews, even if we work totally different jobs: What did you accomplish in the last year? What are your opportunities for improvement? But here's one you might not have seen before: How did you use generative AI at work?
Something like that question might be on the next performance reviews for at least one employer. In a memo posted online after it leaked and was reported on by CNBC and others, Shopify CEO Tobi Lutke said using AI in the workplace is no longer optional at the e-commerce software firm, which employed about 8,100 people at the end of 2024.
"Using AI effectively is now a fundamental expectation of everyone at Shopify," Lutke wrote in the memo.
Gen AI tools like OpenAI's ChatGPT and Google's Gemini are increasingly being touted as game-changers at the office, with business leaders saying they can make employees more efficient. At the same time, that transformation has raised concerns that these tools will replace humans, leading to fewer jobs. A recent Pew Research Center survey found 64% of American adults expected AI growth would lead to fewer jobs.
Shopify is one company emphasizing gen AI in the workplace, but it isn't the only one. What happens when your boss adds "use AI" to your job responsibilities?
Will AI at work lead to fewer jobs?
Lutke's memo emphasized the importance of Shopify's employees tinkering with AI and spelled out certain requirements, including sharing what they've learned about using AI tools. He also said teams would need to demonstrate why AI can't meet needs before asking for more resources or to hire new employees.
The memo clearly shows one potential impact of gen AI on the availability of jobs: Companies will be less willing to hire if that work can be done by AI instead.
That fear is widely shared, with more Americans worried than hopeful about AI's impact on jobs, according to a separate Pew survey released in February that focused on Americans' thoughts on AI at work.
Despite the widespread fears, Nicole Sahin, CEO and founder of G-P, a global employment and human relations firm, told me she still sees companies hiring workers in line with what would be expected in a growing labor market.
"Companies are definitely hiring people and they can't find enough talent," she said. "I don't feel that hiring is slowing down."
What is changing, perhaps, is that the people who are being hired for the kinds of jobs that can be done alongside gen AI tools are being hired based on their ability to be creative and versatile with that technology, Sahin said.
When AI is an expectation at work
The Shopify memo and its expectation around AI use is "the beginning of the new normal," Sahin said. G-P released a survey this week of more than 3,000 global executives and HR professionals, with 91% of executives reporting they're scaling up AI efforts at their companies.
Sahin said she sees the issue as one where companies expect workers to be willing to experiment and be creative with technology. "The willingness to be nimble is extremely important," she said.
Experts say the expanding use of gen AI in the workplace is changing the skills employees need to thrive. Many workers, including those in entry-level positions, will need to rely more on subject matter expertise and judgment rather than the skills to do tasks that can be done by an AI tool instead.
Most workers in the February Pew survey said they don't use AI chatbots at all or use them rarely, and only 16% reported using AI in their jobs.
Even younger workers generally aren't using AI in their jobs. A Gallup survey released this week asked Gen Z adults about their use of gen AI in the workplace. Only 30% said they used it for work, and more than half said their workplace didn't have a formal AI policy. The survey found 29% said AI doesn't exist for their work and 36% said the risks outweighed the benefits in their jobs.
Just because you can or do use AI at work doesn't mean it's worth it. A report this month by the consultancy firm Coastal found half of the business leaders it surveyed said they've seen no measurable return on investment from AI, and only 21% reported proven outcomes. Coastal attributed this gap between hype and results to the disconnect between experimentation and strategy.
"Without clear business alignment or defined outcomes, AI risks staying stuck in the 'interesting but isolated' category," the Coastal report said.
The problem with AI at work
Gen AI systems like ChatGPT may be able to generate answers to a wide variety of queries, but they aren't answering those things the same way a human would. For one, they're prone to errors known as hallucinations -- essentially making stuff up instead of acknowledging they don't know the answer.
That makes it essential to use AI wisely and not trust its answers as always being correct. Especially large, general language models like ChatGPT, which are trained on a vast amount of data, not all of it good or relevant to your job.
Those kinds of models "really should not be used for work," Sahin said. "When you're thinking about using AI in business, it can't hallucinate, it can't get things wrong."
In the workplace, you want specialized tools that are less likely to hallucinate and are easier to verify and correct, she said. Workers need to be able to detect those issues and fix them in order to use AI well.
At Shopify, learning those skills is just part of the job now, Lutke wrote. "Frankly, I don't think it's feasible to opt out of learning the skill of applying AI in your craft; you are welcome to try, but I want to be honest I cannot see this working out today, and definitely not tomorrow."