Shopify CEO Tobi Lütke wrote earlier this year in a company memo that managers must prove that they “cannot get what they want done using AI” before asking for a human headcount or resources.
But the e-commerce giant isn’t alone in shifting to an AI-over-human hiring model—Duolingo’s CEO, Luis von Ahn, also faced backlash on LinkedIn. Users threatened to cancel their subscriptions after he posted a memo about the company’s plans to “gradually stop using contractors to do work that AI can handle.”
Companies have good reasons for their AI enthusiasm:
As companies pursue the promise of doing more with less using AI, more managers must now demonstrate why they require costly human expertise over AI tools. These memos and the “prove AI can’t do it” mentality are a new reality that managers everywhere must face. But a survey from Pluralsight found that 91% of executives are faking their AI knowledge—so in some cases, senior leaders and executives may be asking for technology solutions that they may not fully understand.
So how can managers prove that human solutions are still needed to their AI-bullish bosses? Below are four suggestions to help you make your case.
1. Focus on nuance and complexity
When justifying a human hire, break down what “make AI do it” actually means.
Minyang Jiang, chief strategy officer at Credibly, says it’s vital to ask for critical questions and clarifications in this process. “Do you mean have AI do it end-to-end and not have anyone look at it?” she says. “In which case, how do you do things like quality control?”
Context also matters. “Are you looking at use cases where it’s minutia work-intensive, in-house? Or is it going out to the customer, to the media or to the board?” she continues.
For instance, a small in-house project can probably be done with AI—but AI quality can vary widely, so she adds that humans are still critical in the quality assurance process.
2. Highlight uniquely human skills
“AI is really good at summarizing… [and analyzing] complex documents and explaining things. It’s really good in low-risk scenarios,” Jiang explains. “But it’s not great when things break and you need an expert to really understand and diagnose what’s going on, especially in what they call ’in the border’ or fringe situations.” These one-off situations where tough calls must be made are better left in human hands.
AI also doesn’t have the real, lived experience of being a part of the workforce and relating to other humans. Many workplace challenges require nuance and reading between the lines—and these are skills that AI can struggle with.
“Probably 50% of my time is spent in conflict resolution, de-escalation, aligning people [and] active listening,” Jiang shares. These are skills that particularly challenge AI’s capabilities. “AI is never going to tell you what someone is not saying to you,” she continues. “It’s never going to tell you how to read between the lines…. It’s not going to tell you—at least not today—how to read body language. Humans are actually wired to take in a lot more data from all different kinds of cues.”
3. Address security and trust risks
Beyond soft skills, though, human oversight also plays a critical role in risk management.
“You… have all these third parties that build on top of these large language models… [and] all of those third parties have different data privacy policies,” Jiang says. This could leave a company open to its data being resold or reused to further train tools, even if the tool they’re built on doesn’t allow for that.
The trust gap poses another challenge: “Just because [your boss] thinks [AI] can be trusted… doesn’t mean that your customers and stakeholders actually will trust it,” Jiang observes. “That gap already is fundamental.”
Klarna, a payment company that made headlines for replacing 700 customer service workers with AI, quietly began recruiting humans again after customers complained about decreased customer support quality. Oftentimes, humans want to talk to other humans, not bots that may not be able to answer their questions.
4. Build AI competency, not competition
Instead of positioning the conversation as AI versus humans, demonstrate how humans and AI work best as collaborators. “You need people who have been trained and know how to work with AI to figure out how to critique it,” Jiang says.
Many employees aren’t skilled AI collaborators, though, since most companies have been lagging in workforce training. Only 52% of companies have trained their workforce how to use AI tools, and nearly 75% of employees blame their workplaces for their weak AI skills because they don’t have access to upskilling training or tools.
But even if AI training is on the table, Jiang warns about overreliance. “If you’re not incubating and training your employees internally… sooner or later your company’s just going to lose the expertise if you’re over relying on AI and no one can actually exert executive control,” she says.
Reassigning value to human employees
Humans bring real-world experience to the table that AI can’t replicate since it simply aggregates publicly available information. This means that human experience, deep knowledge and expertise—much of which lives in our brains—isn’t built into these models.
Because of this, humans continue to play a pivotal role in the workplace. Focusing on human strengths and the skills that are harder for AI to replace will be strong strategies to help you remain ROI-positive in these changing times.
Photo courtesy of chayanuphol/Shutterstock.com
