Picture an American factory around 1900. The old steam engine in the basement has just been ripped out and replaced with something new: an electric dynamo. On paper, this should be the productivity miracle of the century. In practice… nothing happens. Output barely moves. For almost thirty years.
The economist Paul David dug into why, and the answer wasn't the technology. It was the floor plan. Steam power had forced every factory into one shape: a single giant engine driving one long shaft, with all the machinery crowded around it in reach of the belts. When electricity arrived, owners did the obvious thing: swapped the dynamo in where the steam engine used to sit and kept everything else exactly the same. It took a full generation of engineers, comfortable enough with electricity to stop imitating steam, to realize the real unlock wasn't a bigger, cleaner power source. It was that a small motor could sit on every machine, which meant you could finally redesign the whole floor around how work actually flowed. That's when productivity took off — decades after the "electricity shift" had technically already happened.
We are, once again, standing in front of the dynamo. And we keep calling this "the AI shift," as if the story is the new power source — as if a 70-year-old technology (the first neural network paper dates to 1958) just now became the whole plot. It hasn't. The interesting shift isn't the motor. It's the floor plan. And the floor plan has a name: Industry 5.0.
The name we keep reaching for is too small
Calling this moment "the AI shift" mistakes the dynamo for the redesign. It shrinks a systemic transformation of labor, organizational design, and what we collectively value - down to a toolset. And when leaders think they're managing a toolset, they make toolset-sized decisions: unplug the old engine, plug in the new one, keep the floor plan exactly as it was. Automate the task, cut the headcount, ship the chatbot. Then they're surprised when the humans they cut turn out to be the thing the new floor plan actually needed.
The European Commission saw this coming before ChatGPT existed. Back in January 2021; while the world was still mid-pandemic, its Directorate-General for Research and Innovation published a framework called Industry 5.0, built on three pillars: sustainability, human-centricity, and resilience. Its core move was reframing the goal of industry away from pure "jobs and growth" and toward something more durable: prosperity that respects planetary boundaries and puts worker wellbeing at the center, not the edges, of the system. It was, in effect, a values upgrade released three years before most companies started panicking about large language models.
That timing matters. It tells us Industry 5.0 was never a reaction to AI. AI just made it urgent.
What the data says humans are actually worth right now
Here's where it gets interesting, and where the accessible science backs up the poetic claim: in a world flooded with machine intelligence, distinctly human capabilities aren't becoming obsolete. Instead, they are becoming the differentiator. The World Economic Forum's Future of Jobs Report lists AI and data literacy as a top skill for the next five years, which surprises no one. But sitting right alongside it, ranked as equally critical, are creative thinking, resilience and flexibility, and empathy and active listening. Not soft skills. Load-bearing ones.
Two companies stumbled into this lesson the expensive way. Klarna laid off around 2,000 employees in 2024, betting that AI could absorb the work. It rehired. Their own explanation, since widely quoted, is worth sitting with: "AI gives us speed. Talent gives us empathy. The magic happens when the two work together." IKEA took the opposite route from the start — instead of automating its call center jobs away, it retrained those employees as interior design advisors. Automation handled the repetitive queries; humans handled the judgment calls. The result was reportedly €1.3 billion in additional revenue. Same technology, opposite philosophy: one company treated automation as a replacement for people, the other treated it as a liberation of people into more human work.
That's the whole Industry 5.0 thesis in two case studies. The machines don't have to shrink the humans. They can also be the thing that finally frees us to do the parts only we can do.
So what does this mean if you're not running a factory or a bank?
Almost none of us are managing assembly lines. But every team, studio, and organization is currently deciding — often by accident, often under deadline pressure — whether it's building an AI-shift company or an Industry-5.0 company. The first optimizes for speed and headcount reduction and quietly hopes the humans won't notice they're being designed out. The second optimizes for what humans and machines can do together that neither can do alone, and treats resilience and wellbeing as strategy, not HR line items.
This is a genuinely rehearsable choice, not just a philosophical one. Organizations that get this right aren't the ones with the best prompt engineers. They're the ones who've actually imagined, in detail, what a good version of this transition looks like for their people, and then built toward it on purpose. And the imagination-step is far more important than you might think. It is crucial to spend time there. When we vividly imagine a future state, we activate the same brain networks used for remembering the past — we create something close to a memory of a future that hasn't happened yet, and that "memory" quietly shapes the decisions we make on the way there (more about this in our article The Psychology of Imagination). Skip the imagining, and you default to the anxious version by inertia. Do the imagining deliberately — through scenario work, prototypes, speculative design — and you give your organization something to walk toward instead of something to flee.
The practical version of "imagine it first"
Before you write the AI policy, run the workshop. Before you announce the automation roadmap, prototype what the human side of the new workflow actually feels like for the people in it. Not in a slide, but in something they can walk through.
Ask the Industry 5.0 questions explicitly:
Where does this make us more resilient, not just faster?
Where does it free judgment and creativity to do more, rather than freeing them from the org chart entirely?
Ask your people what they wish they could do more of, what ideas they have that they never have time for. Ask what they wish somebody else did and where their unique creativity, their social qualities and human presence could make a difference. Ask them what they wish they could change. Ask them what they would do if they had a team of assistants but kept their role and responsibilities. What would they still do and what would they do instead or more of if they could outsource anything they wanted? How would that change everything?
Those questions are cheap to ask in a workshop and extraordinarily expensive to ask after the layoffs.
Your imagination is the first draft of tomorrow: so draft with intention! Right now, a lot of organizations are drafting a future where humans are optional, the way that first generation of factory owners drafted a future that was just steam power with better wiring. History, including the one being written in real time by companies like Klarna, suggests that's a bad draft, and a slow one. The better one is available. It just has to be imagined on purpose, by the people who'll have to live in it.
So: are you building the AI shift, or the Industry 5.0 one? And who in your organization is actually being asked?
Sources
Read more
