Peter Poon

The AI Paradox: Rising Jobs, Falling Employment We’re seeing a clear divergence across the data: AI-related job postings are accelerating, yet overall employment levels are softening. This paradox highlights a deeper issue with the design of labour market metrics. Government planning models, fiscal allocations, and productivity forecasts still rely on traditional employment indicators, assuming that job creation equates to employment growth. But AI-driven roles don’t scale that way. • One AI engineer can replace multiple operational roles. • Many AI jobs are high-skill, low-volume, clustered in a few sectors and regions. • Automation, synthetic data generation, and algorithmic productivity are inflating output, without a corresponding rise in headcount. This creates blind spots in how governments and institutions allocate funding, model tax receipts, and plan for skills investment. The “missing employment” is not necessarily economic weakness; it’s a structural data lag. To keep pace, we need modernised metrics that capture: • Digital and augmented productivity • AI’s contribution to economic output • The redistribution of labour from human to algorithmic functions Without this, policy risks being anchored to an outdated understanding of work; underestimating both the value created and the displacement underway. The labour market isn’t collapsing. It’s evolving faster than our ability to measure it. #AI #FutureOfWork #Economics #Policy #Innovation #LabourMarket #Automation #DataEconomy

10

Jane Goodall started studying chimpanzees in 1960 with a pencil and a notebook. Over the next 60 years, the Jane Goodall Institute accumulated one of the world's most comprehensive datasets on primate behavior—thousands of hours of footage, photographs, and handwritten field observations. But there was a problem. Most of this knowledge was effectively locked away. Researchers who needed to review specific behaviors had to manually flip through pages or physically pull files from cabinets. A single query could take days or weeks. Now, they're using AI to unlock it all. (You can read more in "How the Jane Goodall Institute uses AI to develop a better understanding of chimpanzee behavior" at https://bit.ly/3JmrNum) The Institute is training computer vision models to identify individual chimpanzees across decades of footage and automatically detect specific behaviors like tool use. What took researchers weeks now happens in minutes. I keep thinking about the parallels to enterprise organizations. Most companies I work with have their own version of this problem. Decades of customer interactions buried in old CRM systems. Engineering knowledge locked in the heads of employees who retired years ago. Product insights scattered across departments that don't talk to each other. It's all technically "there"—but functionally invisible when your teams need it. The breakthrough isn't just about digitizing old records. It's about using AI to surface patterns and insights that would be impossible to find manually. The Jane Goodall Institute isn't just making their data searchable—they're making it reveal things they never knew to look for. So here's my question for leaders: What's your organization's filing cabinet? Where is the institutional knowledge that could transform how you serve customers, develop products, or make decisions—if only you could access it at scale? The technology to unlock it exists today. The harder question is figuring out where to start—and I'd be curious to hear how other leaders are thinking about this. #AWSExecutiveInResidence #DigitalTransformation #AI #EnterpriseStrategy

25

2 Comments