Leaders Accept AI’s Slow Impact on Work

Since the introduction of ChatGPT, businesses have been looking forward to the day when AI will empower their employees and transform their businesses forever. Three years later, they are still waiting. For what? And what is the solution?

These two questions dominated almost every conversation I had with executives in Davos last week — including a Business Insider roundtable I moderated with 15 HR directors and other senior executives.

One explanation that kept coming up was incomplete adoption by employees. Many professionals are rightly concerned about the impact of these tools on their work, or at least skeptical of their usefulness, because AI waste abounds. To overcome these hesitations, bosses have increased the pressure, making the use of AI compulsory and integrating it into performance evaluations.

But a number of leaders at the roundtable advised against resorting to strong-arm measures. Cisco learned this the hard way. “When we required our employees to undergo mandatory AI training, not only did it not promote sustainable use, it actually had a slightly negative impact,” said Francine Katsoudas, the company’s director of people, policy and purpose. What worked, she explained, was “giving choice,” like when Cisco gave its engineers access to a half-dozen different AI tools, allowing them to decide which ones to use and how to use them. “They really loved it,” she said.

Another theory was that even if employees to want To use these new tools, they do not have the skills to get the most out of them. Part of the solution, some say, is hiring people who are already well-versed in AI. Kyle Lutnick, executive vice president of Cantor Fitzgerald, said he wants to recruit more new college graduates at a time when other companies are hiring fewer of them, precisely because they are more proficient with these tools than their older counterparts. But recruiting new blood will not be enough. Employers will also need to do much more to train their existing workforce. “Investments have been primarily focused on technology and not so much on people,” said Elizabeth Faber, global director of people and purpose at Deloitte. “This needs to change.”

A third explanation is that significant productivity gains require a fundamental overhaul of how work is done within companies. If the Googles or Amazons of the world were to start from scratch today, they certainly wouldn’t have the team structures, workflows, and job descriptions they have now. I think this is why we are most clearly seeing the AI ​​revolution right now in early-stage startups, which are starting from scratch in the post-ChatGPT era.

“84% of work processes were left in their original state during AI adoption and were not redesigned,” Faber said. “So 16% of organizations and work processes are actually natively developed by AI.”

All these proposed solutions are far from being miracle solutions. Encouraging employees to join voluntarily takes more time than threatening to fire them. Training staff – and making them learn – also takes time.

Redesigning tasks will prove to be an even bigger task. Many large companies don’t even know what their employees do on a daily basis. It’s painstaking work to build a comprehensive database of employee skills and the tasks they perform, then systematically determine which of them can be delegated to AI and which cannot. One HR manager I spoke to said it would take years for his HR department to complete this process across all functions of his company.

Once all this hard work is done, what will these companies look like? I posed the question to the group during the roundtable, asking how many of them expect their numbers to decline within three to five years. Two out of 15 raised their hands – I suspect the number would have been higher if I hadn’t been there. One of them, Gina Vargiu-Breuer, director of human resources at SAP, explained that her company is currently keeping its workforce stable as business continues to grow.

“But when you’re not growing, I think that’s when you have to ask, ‘OK, should we downsize?'” she said. “I have a lot of peers at German companies where they’re starting to significantly downsize. So it’s a reality. For us it’s not the case because we’re growing, but I think it will happen in the future.”

Even at SAP? I asked.

“Right now we’re staying flat,” she said. “But if productivity increases and growth slows, then I think we need to look at the situation with a different eye.”

At the end of the week, I left Davos feeling that a workplace truly reshaped by AI – one that would allow businesses to operate with significantly smaller teams – would not arrive as quickly as I had imagined. Those days still seem a long way off, given all the painstaking work companies have to do to get there.

This is something that many economists predicted early on, given how difficult it has always been to fully integrate new technologies into the workplace. They told me that things would change less than expected in the short term, and much more in the long term. No matter how quickly technology advances, humans change less easily.

Leaders around the world are realizing the same thing, which is probably why I detected some frustration in Davos. Svenja Gudell, Indeed’s chief economist, likened the global emergency around AI to the impatience of a parent potty training their child. “It’s complicated, it’s a long process,” she said. “You ask yourself, ‘Why isn’t this happening? It’s been three weeks already.'” His message to leaders: “Give yourself some grace.”

The slower timetable is good news for the rest of us: It gives us time to learn new skills, debate new public policies, and try to shape the future we actually want. But it would be a mistake to take the modest changes so far as evidence that tectonic shifts will not occur. I came away from Davos convinced that when they happen, they will be far greater than anything we imagine today, for better or for worse.

Aki Ito is chief correspondent for Business Insider.