The Partnership Imperative, a perspective shift.
Mar 15, 2026
How AI Is Reshaping Skills, Work, and Human Value
The question dominating boardrooms and policy halls alike used to be whether artificial intelligence will transform work, just under a year ago. “AI is going to steal people’s jobs”, was the common outcry.
Now it has shifted to how — and for whose benefit.
A growing body of research from institutions ranging from McKinsey to MIT is converging on a nuanced answer: the future of work is not a story of replacement, but of partnership.
The challenge is ensuring that partnership is designed deliberately, rather than left to market forces alone.
The scale of disruption is overwhelming. The World Economic Forum's Future of Jobs Report 2025, drawing on surveys of over 1,000 major employers representing 14 million workers, projects that 92 million roles will be displaced by technological and macroeconomic shifts by 2030 — yet 170 million new jobs will simultaneously be created, yielding a net gain of 78 million positions.
What is less certain is the quality and distribution of those new roles, and whether the workforce will be equipped to fill them. The same report finds that 39% of the skills currently required in the job market will change by 2030, with AI and data literacy topping the list of capabilities employers are urgently seeking.
McKinsey's Agents, Robots, and Us (2025) adds more clarity to this picture. Its analysis suggests that approximately 44% of work hours could eventually be handled by AI agents, with a further 13% by physical robots — leaving a resilient 43% squarely in human hands. Critically, the report argues that the greatest economic value, estimated at $2.9 trillion annually in the US alone, will come from redesigning entire workflows end-to-end.
This is the key. Redesigning workflows.
Try to imagine for a second what this actually means. The challenges, but also the opportunities.
There is a distinction between automation and augmentation.
And it sits at the heart of new academic work from MIT.
Research by professors Roberto Rigobon and Isabella Loaiza introduces the EPOCH framework, identifying five categories of uniquely human capability that AI consistently struggles to replicate: Empathy and Emotional Intelligence, Presence and Networking, Opinion, Judgment, and Ethics, Creativity and Imagination, and Hope, Vision, and Leadership.[1] Their analysis of the US labor market between 2016 and 2024 found that human-intensive tasks — those requiring high EPOCH capabilities — have actually increased in prevalence, and that new tasks added to the workforce in 2024 carry higher EPOCH scores than those that disappeared. "We deliberately don't call these 'soft' skills," notes Rigobon. "A 'hard' skill, like solving a math problem, is comparatively easy to teach. It is much harder to teach a person empathy, creativity, or hope."
The policy dimension of this debate has been sharpened by economists Daron Acemoglu, David Autor, and Simon Johnson, whose 2026 Brookings paper, Building Pro-Worker AI, draws a critical distinction between technologies that automate expertise and those that amplify it.
They argue that current AI investment is disproportionately tilted toward automation and the pursuit of artificial general intelligence, with comparatively little directed toward tools that expand what workers can do.
Their framework identifies "new task-creating technologies" as unambiguously pro-worker — generating demand for novel human expertise rather than substituting for existing skills. The implication for organizations is direct: the design choices made today about how AI is deployed will determine whether it widens or narrows economic opportunity.
Not an easy proposition to contend with.
What emerges from this convergence of research is a clear imperative.
AI fluency, which is the ability to work alongside, direct, and critically evaluate intelligent systems, is rapidly becoming a foundational skill across every sector, not merely in technology.
The WEF reports that two-thirds of employers now plan to hire specifically for AI-related capabilities. Harvard Business School research further shows that workers collaborating with AI are significantly more likely to produce ideas in the top 10% of quality rankings, suggesting that the human-AI team, when well-designed, outperforms either party working alone.
The transition will not be frictionless.
It demands that organizations stop treating AI as a cost-reduction tool applied to existing processes, and start treating it as a catalyst for reimagining what work can be.
It demands that education systems embed AI literacy early and broadly.
And it demands, as Acemoglu and his colleagues argue, that the institutions shaping AI development be held accountable for building technology that makes human skills more valuable, not less.
The machines are capable. The question is whether our strategies, our institutions, and our leadership are versatile and rightly inspired to handle this disruption.
