Dr Fazal Ali
AI is a foundational layer of labour. It replaces, augments, and manages human workloads. Digital workers and AI agents autonomously handle multistep administrative tasks, navigate applications, collect data, fill out forms, and trigger workflows to deliver client-facing services with extreme velocity. Workers are completing tasks with greater agility and efficiency. But economies don’t reflect this efficiency.
Two curious things are happening in the Intelligent Age. Economic expansion is still advancing, while job growth slows to a crawl. This indicates that workforce productivity is increasing. But many measures of productivity growth have remained fairly stable. Typically, these things are incongruent or cannot be true simultaneously. Could AI be specifically responsible for this curious productivity paradox?
There are two primary metrics used to gauge productivity, and the two are pointing in completely opposite directions. One is labour productivity, which measures output per unit of labour. The other is total factor productivity (TFP), a broader metric that encompasses how efficiently the entire economy converts input into output.
While the global economy is expected to expand overall, this growth is often described as modest and fragile rather than strong and widespread. The IMF’s latest forecast projects a global real GDP increase of around 3.1 per cent in 2026, while the OECD remains more cautious at approximately 2.8 per cent. Both organisations acknowledge that the momentum is slowing compared to the vigorous post-pandemic rebound.
Crucially, this expansion is uneven: emerging and developing economies are typically growing faster than advanced economies. This means the global economy is still advancing, but the key driver is the faster growth in emerging regions, rather than a uniform, synchronised boom.
Emerging markets and developing economies (EMDEs) are outpacing advanced economies and now account for about 60 per cent of global GDP in purchasing power parity (PPP) terms. This is fuelling a multi-speed global expansion in the Age of AI.
A key feature of the AI productivity paradox is the “botsitting effect”. AI agents can manage entire workflows independently. They are routinely used to triage emails, analyse market data, edit videos, manage e-commerce stores, and handle customer service.
Rather than handling daily operations, entrepreneurs act more like business orchestrators. Entrepreneurs train their agents by refining instructions and approving automation protocols, effectively letting the AI do the heavy lifting while they focus on high-level strategy.
The phenomenon has become a cultural and economic craze in China, with influencers claiming that failing to “raise” an agent or “claw” means falling behind. Local governments in tech hubs like Shenzhen and Suzhou are actively offering incentives to SMEs to boost these companies.
The intense, continuous processing required to run these agents can be expensive. Additionally, Chinese cybersecurity experts have warned that granting AI agents such deep access to computer files and web browsers poses significant risks.
This shift to “lobster-raising” has driven a major evolution in Chinese entrepreneurship, enabling a single individual to run the equivalent of an entire corporate team with the help of AI. In Europe, while white-collar workers save considerable time with AI Agents, they then spend an average of 6.4 to 11 hours per week managing, feeding context, and debugging the models, erasing aggregate time savings.
Huge corporate investments in AI infrastructure, such as data centres and hardware, are actively supercharging GDP growth and economic expansion. However, these capital investments take time to translate into measurable “output per hour” for individual workers.
While many employees use AI, a smaller fraction utilise it frequently enough to significantly impact total firm performance, creating localised pockets of productivity that don’t shift national data. Adoption remains shallow. In some jobs, workers who use AI are more likely to produce the same amount of work in less time.
It is estimated that the effect is the saving of one workday per week. This capital deepening occurs when workers have access to better tools. Their individual productivity rises as a result. In the past, this happened when a construction worker traded in a shovel for an excavator.
During the early and mid-1990s, when the Internet was in its infancy, workers had unfettered access to groundbreaking technology, but many firms remained stuck in the trenches of a “productivity paradox” as huge IT investments failed to translate into improved efficiency.
That lull did not persist, and if history repeats itself, we might be in the early days of another historic productivity surge without even realising it. But the growing pains come with social costs.
Workers who use AI tools do save time on their tasks, but that time is often redirected into other work, resulting in fewer breaks overall. The result is more time on the job for most workers, and a higher risk of burnout. Extensive use of AI at work could also lead to excessive cognitive load.
Dr Fazal Ali completed his Master's in Philosophy at the University of the West Indies. He was a Commonwealth Scholar who attended the University of Cambridge, Hughes Hall; the Provost of the University of Trinidad and Tobago; the acting President of UTT; and the Chairman of the Teaching Service Commission. He is the President of NIHERST and an external services consultant with the IDB.
