To paraphrase Robert Solow and his eponymous paradox, ”You can see AI everywhere but in the productivity statistics”.
The Solow paradox described the slowdown in overall productivity growth despite rapid technological advancement.
It turns out that this has paradox has remained true through the technology adoption waves of computing, ERP, internet, cloud and even AI.
A recent Economist report “Solow, the sequel” comments that “Artificial Intelligence is improving fast. Its effect on output, not so much”.
Despite the fact that AI is advancing at startling speed, it seems that there remains a deep flaw in the argument that AI is powering a productivity boom.
The Economist article reports;
- “Such improvements are usually made not just when workers use a new tool more often, but also when firms reorganise production around it. Early factories became only a bit more efficient when steam engines were replaced with electric motors; the real revolution came decades later after floor plans were redesigned to make the most of electric power. More recently, productivity growth was a disappointment for years after personal computers became widespread. It sped up only once firms adopted business models that exploited the technology to its full potential. Much of America’s productivity revival in the 1990s came not from Silicon Valley but from retail, where computers transformed logistics and inventory management.”
Some observers point to the idea that AI-driven layoffs are indicators of productivity improvements.
- The Washington Post produced an interesting analysis “Layoffs might feel like they’re soaring, but they’re not”. High-profile companies, including Meta and Uber, have effectively said they’re spending so much money on AI that they need to cut back elsewhere, including by slashing staff or reducing planned hiring.
- That’s a very different thing . . . .
Fortune reports that the CEOs of OpenAI and Anthropic, who spent the last year warning that AI would obliterate white-collar employment, now admit they were wrong, joining other leaders like Goldman Sachs CEO David Solomon in casting doubt on an AI job apocalypse.
Getting to the foundations of productivity, Glen McCracken shared a great analysis of the true impact of automation;
- We tend to think that automation is mainly about labour reduction:
- reduce cost, remove steps, increase efficiency, move faster
- What automation really does is reduce the time between intention and consequence.
- Things happen faster: retrieval, routing, approvals, recommendations, communication, execution
- Sounds great … until the underlying logic or assumptions are shown to be flawed.
- Automation compresses the speed of failure.
- A bad process with humans often breaks slowly enough for somebody to notice:
- An experienced operator intervenes, a manager questions the output, a delay accidentally creates time for reflection
- Human friction often acts as a safety mechanism. Ironically, inefficiency can create resilience.
- With AI, a flawed assumption can spread instantly across workflows, systems and decisions before anybody fully understands what is happening.
- “Operational amplification”.
Glen nails it with “what automation really does is reduce the time between intention and consequence”.
This makes the productivity equation more complex.
Whilst we continue to throw the latest technology (AI, in this case) at small collections of tasks, we continue to miss the point.
Genuine economic productivity opportunities exist in the end-to-end cycles of business. You will hear me continually “bang on” about the need for “end-to-end process” enablement. This is synonymous with “Systems Thinking”, approaching a problem by looking at the overall “system” (a.k.a business process), not just it’s individual parts;
- Examining the whole (end to end process) rather than just the parts (tasks)
- Looking at the feedback loops of action, reaction and result
- Recognising that a change to one part (or task) in the “system” (process) will indirectly or directly affect other elements of the process, upstream and downstream.
This, to me, gets to the root of the issue that “improvements are usually made not just when workers use a new tool more often, but also when firms reorganise production around it.”
That is a manifesto for “end to end process enablement”, if ever I heard one 😉
You can see AI everywhere but in the productivity statistics.
There is little sign that AI has reached a stage of measurable business impact and economic productivity, yet. Nine out of ten senior managers see no measurable improvement in labour productivity.
The organisational rewiring, in other words, has barely begun. Something big may indeed be happening with AI itself.
For now, it remains largely invisible in the macroeconomic data.
You can read the Economist report “The AI productivity boom is not here (yet)” here . . .
You can read Glen McCracken’s post “Automation compresses latency” here . .
You can see AI everywhere but in the productivity statistics
Thanks for reading . . .