MIT Dropped Some Serious Research into Where AI Can Realistically Be Applied Across Human Work!

WinMax Blog Team

14 July 2026
0 Comments

On 21st March 2026, a research team headed by Professor Thomas Malone at MIT's Center for Collective Intelligence released a report - Where can AI be used? Insights from a deep ontology of work activities. - Taking a bottom-up approach to the relationship of work, and its viability to be processed by AI or Robotics, the paper analyses tasks, and their propensity to realistically attract automation as a replacement to human executed work. Instead of analyzing jobs at the occupation level, the authors break work into nearly 40,000 detailed activities and map real-world AI usage onto them using data from 13.275 AI software applications and 20.8 million robots worldwide, developing a comprehensive framework for understanding where AI can realistically be applied across human work activities. 

The MaxAITM Daily Report from WinMax, will dissect and inspect this highly scientific, and pragmatic report in a series of blogposts starting this week. As a Tech Staffing & Consulting Company located in the Bay Area over the past 20 years, we appreciate this timely report, which in place of alarmist headlines-from-an-headless-chicken, and the weak-tea-analysis we commonly hear online like "Elon Says These 5 roles Will be Non-existent in 5 years" or "Satya Says Those Roles Most in Danger of...", categorically lays out a realistic deep-dive into the nature of work, not analyzing jobs at an occupation level, but breaking it down into nearly 40,000 detailed activities. The report then builds it back up, mapping real-world AI usage onto these activities. We can also extrapolate from it, the crucial distinctions between B2C AI and B2B AI. Currently, most understanding of the AI capabilities, and its utilities; is homogeneous, with no distinction made between vague prompts about the disposition of cholic cats vs a series of algorithmic prompts that could lead to a hedge-fund grade recommendation report. A single Google Search of AI generated art or images is enough to assume that human beings are largely untalented, and no amount of AI support can bring sublimity or meaning to an artwork unless the prompter has aesthetics and design sensibilities. It just works to the advantage of the industry to club the stats, so it looks like a torrent of generic disruption, instead of the actuality which is a siloed usage across general and business.

Please tune into The MaxAITM Daily Report over the next couple of weeks as we deep-dive into the MIT Report, and see how they map to skills and staffing management in todays business space. Please read this post further for the highlights and major findings of the report. You can read the MIT Report in its entirety here 

Framing the Question: Which jobs will AI replace? vs Which activities inside those jobs can AI perform? Which 10%, 30%, or 60% of activities inside that role can be improved by AI?

The Premise

The researchers argue that jobs are too broad and too complicated to be the unit of analysis. Understanding AI's impact on professions requires dissecting jobs into their component activities. Only then can we accurately evaluate where AI can genuinely create value. Instead of seeing AI as a force that automates entire professions, the researchers show that AI is far better understood as a technology that targets specific activities—writing, analyzing, summarizing, communicating, predicting, generating, coding, designing, and thousands of others. The findings are both exciting and deeply practical for entrepreneurs, executives, consultants, technology vendors, and professionals trying to understand where the AI train is headed

The First Principles

Staring with O*NET, the U.S. Department of Labor's occupational database containing approximately 20,000 work tasks, the researchers broke tasks into smaller atomic activities, reorganized and standardized them, built hierarchical relationships between activities, and created a framework containing nearly 40,000 activities. This deep analysis reveals that at the highest level, everything humans do falls into the following three categories:

  • Think

  • Do

  • Interact

Takeaway #1: AI Adoption Is  Concentrated 

The top 1.6% of work activities capture more than 60% of total AI market value. AI is not evenly distributed across work. It creates extraordinary value in a relatively small number of activities. Most companies launch broad AI programs intended to transform everything simultaneously. The report suggests a much smarter approach - Find the handful of activities where AI performs exceptionally well. Start there. Maximize gains. Expand later. The companies achieving the greatest AI ROI are rarely those implementing AI everywhere. They are usually those implementing AI precisely where it matters most.

Takeaway #2: Opportunities are Hidden Yet

The valuable insight in the paper is not just on where AI is being used. It's also on where AI is not being used. It reveals hundreds of activities with little or no meaningful AI adoption today. Some of these activities are difficult because they require intangible aspects like empathy, trust, human presence, social intelligence etc., while others simply haven't attracted enough entrepreneurial attention yet. Not every gap represents a limitation though. Some gaps represent opportunity. Founders, product managers, or entrepreneurs, would do well not only to look where AI is already succeeding, but also where no one has built a solution yet that solves underserved activities hidden deep within business workflows. 

Takeaway #3: The Future of Work Is Granular

For decades, economists have discussed technological disruption at the occupation level. The MIT study suggests a different future. Instead of jobs disappearing wholesale, we are bound to see thousands of individual activities transformed. Some activities will be almost completely automated. Others will be partially augmented. Some may remain predominantly human for decades. The accountant of 2030 may spend far less time reconciling transactions and far more time advising clients. The salesperson of 2030 may spend less time creating proposals and more time building relationships. The manager of 2030 may spend less time reporting status and more time making decisions. The roles survives. The activity mix changes.

Takeaway #4: AI as The Great Work Augmenter

AI is not a general-purpose replacement for human work. It is better understood as a highly capable tool for specific categories of activities. The big takeaway from the report is that AI is less a technology that replaces jobs, and more a technology that reshapes activities. And the organizations that learn to map, measure, and optimize these activities will gain the greatest advantage in the AI Future.

For more learnings and deeper drill-down into the MIT Report, tune into The MaxAITM Daily Report on our website www.winmaxcorp.com 

WinMax Blog Team

Founded in 2005, WinMax is an IT recruiting firm that specializes in connecting the best talents from the tech world to the right organization. As technology recruiters, we can help you find the right IT personnel possessing the skills, expertise, and values that reflect your organizational needs and goals.

pexels-djordje-petrovic-2102416

Talent Optimized!

As technology talent acquisition specialists, we can help you find the right technology talent for your projects.

Comments

Subscribe to our Blog