Imagine the floor of an auto manufacturing facility in the 1980s. A line worker endures the daily grind of repeating the same 45-second task, such as bolting two fenders or clipping a wire harness. A line worker’s suggestions on how to fix bottlenecks or improve the flow are often ignored. Making the rounds is a foreman checking attendance, enforcing a strict pace, monitoring quality, and filing end-of-shift reports on output and breakdowns.
Fast forward to the mid-1990s. The deployment of advanced technologies like robotics, programmable logic controllers (PLCs), and manufacturing execution systems (MES) is reshaping how work gets done. There are fewer line workers on the floor handling a broader set of tasks, like clearing jams and loading robotic cells, while producing twice the output. Supervisors are now empowered with production data to coordinate across automated processes and cross-functional teams.
Over this century of progress, these new systems required specialized expertise to operate and optimize. That expertise sat far from the production line, concentrating value among engineers, analysts, and supervisory roles. Supervisory roles saw wages grow much faster than the pay of production workers.
The value of these technologies didn’t trickle down to line workers. From the 1980s to the 2000s, new knowledge layers amplified supervisors, increasing the wage disparity between management and production staff. This is a common scenario across industries as a result of historical technological disruptions.
According to data from the Bureau of Labor Statistics, waves of automation and digital transformation have added more than 70 million net new jobs to the U.S. economy since 1980. Yet productivity has grown 2.7 times more than average compensation (wages and benefits) for production and nonsupervisory workers since 1979, according to the Economic Policy Institute. This group represents roughly 80% of the U.S. workforce.
AI could flip the script because it is the first technology that is accessible through natural language rather than specialized technical skills. Past technological disruptions required workers to learn software, programming, or analytical tools to benefit from new systems. This created a push factor that required businesses to lead adoption and integration. AI lowers the entry barrier with tools that extend and amplify human capability. The interface is a conversation rather than code.
As a result, consumer AI apps like ChatGPT, Gemini, Claude and CapCut are helping drive adoption. And yet, businesses are struggling to create tangible, scaled value that justifies the billions of dollars of investment in this technology.
Over 80% of AI projects fail to deliver business value, with 84% of those failures tied to leadership gaps, like unclear metrics, underinvestment, and unfocused sponsorship, according to research from Pertama Partners. Deploying generic machine intelligence inside legacy structures, without accounting for the real, often messy ways teams actually work, is a recipe for failure.
AI works best when it’s built into the work people do. Bringing the context of workers’ lived realities and frontline insight into systems and structures redesigned with AI is a core differentiator. Technicians, operators, nurses, and service workers often understand the nuances of real systems better than anyone else. This contextual dynamic creates a pull factor, drawing AI-enabled decision-making and value creation toward the edge of organizations.
This is why we predict that intelligence will no longer concentrate in a narrow band of supervisory or white-collar roles. It will diffuse across every layer and role of an organization, reshaping organizational hierarchies and previously exaggerated socio-economic disparities.
Our research estimates that AI can affect 93% of U.S. jobs. Knowledge fields such as law, management, and healthcare have the highest theoretical exposure of 40% to 60%. While blue-collar and manual roles have lower direct automation scores, AI presents a substantial indirect impact as it wraps around physical work.
By amplifying their skills and experience with AI, workers can create hybrid intelligence to produce higher-value output. An HVAC technician can detect early compressor failures for proactive maintenance through AI-driven diagnostics. A bank teller can leverage AI-augmented compliance checks and personalized recommendations for customers. A customer service representative can use AI to gauge customer sentiment in real time, enabling more empathetic responses and faster conflict resolution.
When productivity gains originate closer to the work, workers gain tangible leverage over decisions, processes, and outcomes, opening paths for stronger wage premiums and social upward mobility. This shift in dynamics from previous technology waves puts pressure on legacy organizational and societal systems to adapt.
Current reports warn that AI will replace workers, eliminate entry-level roles, and accelerate jobless growth across industries. But if AI increases worker productivity, there is potential for the technology to rebalance the labor market, give workers more leverage, and improve the circumstances for workers.
The Bureau of Labor Statistics data shows that while unemployment rose to 4.4% in February 2026, aggregate U.S. productivity grew 2.2% in 2025, up from 1.4% in prior years. These gains are likely from investments in tools, data and process redesign rather than per-worker efficiency alone, yet it also serves as an early signal of AI-driven productivity.
AI has the potential to create new models and systems that expand equitable socioeconomic opportunity. Historically, rising output often concentrated value in the hands of a few. But that outcome is not inevitable.
When intelligence moves directly into the hands of people closest to the work, adoption, iteration, and value realization accelerate. Jobs, training, autonomy, and rewards that are thoughtfully designed for human and AI collaboration can distribute value more broadly.
AI is advancing at a pace rarely seen in previous technological revolutions. But the most important innovation of the coming decade may not come from artificial intelligence. It will come from empowering every worker to use it to generate economic and societal value.
Leave a comment
