
What this covers
This is an Economics Explained video built around the MIT Iceberg Index, a tool that maps AI capabilities against human skills for 151 million American workers across 923 occupations and 3,000 counties. The host explains that it works by putting AI and human work on the same skill taxonomy, and that it weights exposure by wage value rather than time spent. The reasoning is that automating 60% of someone's hours does not automate 60% of their value if the remainder is complex judgment. Early on, the video explains why GDP, unemployment and wage data miss task-level disruption, and compares this to the way GDP struggled to value Wikipedia, which led the US Bureau of Economic Analysis to build a Digital Economy Satellite Account.
The middle of the video covers the findings. Exposed work clusters around reading, writing, analyzing and summarizing, and these workers earn more, are more likely to hold graduate degrees, and skew female. The video also reports that conventional state-level metrics explain less than 5% of the variation in index scores, which is why South Dakota, North Carolina and Utah rank above California or Virginia. It gives a gap between theoretical and observed use, with AI able to handle about 94% of computer and math tasks but doing roughly 33% in practice. Examples include IBM trimming HR and Salesforce halting some hiring, along with figures on falling entry-level employment and postings.
The last section turns speculative. About 30% of workers, in physical and relational jobs, have essentially no exposure, and the host applies Baumol's Cost Disease, drawing on the 1965 string-quartet example from William Baumol and William Bowen. The argument is that healthcare, education and care services will become steadily less affordable, with pressure falling on stretched public budgets. The host states plainly that this is an extrapolation rather than a result of the study. He also notes the index leaves out robotics, outcomes and timing, and compares it to an earthquake risk map rather than a weather forecast.
The AI disruption story is fundamentally misframed: most headlines focus on visible tech job losses (2.2% of wage value), but the true exposure lies in cognitive and administrative work across non-tech sectors (11.7% total), creating a hidden crisis in less-discussed states and industries while workers in untouched fields face rising costs from relative productivity declines.
- The Iceberg Index reveals 11.7% of US labor market wage value is AI-exposed, five times larger than the visible 2.2% in tech
- Geographic exposure concentrates in administrative/financial hubs like South Dakota, North Carolina, and Utah rather than tech centers
- Workers in AI-safe industries face accelerating costs due to Baumol's cost disease as productivity surges elsewhere
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Standard economic metrics like GDP, per capita income, and unemployment explain less than 5% of the variation in Iceberg Index scores across states, and in some cases the relationship even flips, meaning states that look safest by conventional measures aren't necessarily the least exposed.
“The tools they're using cannot see this kind of risk. GDP, per capita income, and unemployment, the standard metrics, explain less than 5% of the variation in Iceberg Index scores across states. In some cases, the relationship even flips, meaning states that look safest by conventional measures aren't necessarily the least exposed.”
The Iceberg Index maps AI capabilities against human skills across 151 million American workers in 923 occupations and 3,000 counties, creating an apples-to-apples comparison of technical AI capability and human work using the same skill taxonomy.
“MIT is arguing that the AI era calls for a similar overhaul. Intelligence is now a shared input between humans and machines, and the tools we inherited were simply never designed to measure that. The Iceberg Index is MIT's attempt to build something that is.”
In the internet era, GDP encountered a problem measuring the value of services like Wikipedia that replaced physical products like encyclopedias, forcing the US Bureau of Economic Analysis to create a separate Digital Economy Satellite Account.
“In the internet era, GDP ran into a rather embarrassing problem, which is that it could measure the value of encyclopedias sold in shops, but couldn't measure the value of Wikipedia, which replaced them altogether. So, the US Bureau of Economic Analysis eventually had to build an entirely separate accounting framework, the Digital Economy Satellite Account, just to capture the value of services that existed outside normal market transactions.”
Most headlines and public discourse about AI focus on job replacement, but the more accurate pattern is that AI replaces tasks inside jobs rather than eliminating entire positions.
“Most people, and most headlines, treat AI as something that replaces jobs, but that's not how it works. More often, AI replaces the tasks inside them.”
South Dakota, North Carolina, and Utah show higher AI exposure values than California or Virginia, despite expectations that California would be most exposed.
“According to the Iceberg Index, South Dakota, North Carolina, and Utah show higher exposure values than California or Virginia.”
The most AI-exposed occupations are concentrated in reading, writing, analyzing, and summarizing information, and these workers earn 47% more on average than the least exposed, are nearly four times as likely to hold a graduate degree, and are 16 percentage points more likely to be female.
“the most exposed group earns 47% more on average than the least exposed, is nearly four times as likely to hold a graduate degree, and is 16 percentage points more likely to be female. In plain terms, anyone whose working day is built primarily around reading, writing, analyzing, and summarizing information.”
When the Iceberg Index is applied to the whole economy rather than just tech, the exposure jumps to 11.7% of total US labor market wage value (roughly $1.2 trillion), five times larger than the visible tech sector exposure.
“but when you apply the same methodology to the whole economy, the number jumps to 11.7%, roughly $1.2 trillion, and five times larger.”
The capabilities driving the 2.2% tech sector exposure—document processing, routine analysis, and data handling—appear across hundreds of occupations outside tech including financial analyst, HR coordinator, insurance claims processor, and legal secretary work.
“The same capabilities that make a coding assistant useful to a software engineer overlap heavily with the work of a financial analyst, an HR coordinator, an insurance claims processor, and a legal secretary. So, when you run the index across all the skills, that's where you see the number jump from 2.2% to 11.7% of the total US labor market”
Historical economic transitions that created as many new roles as they disrupted were successful because people could see what was coming clearly enough to prepare.
“transitions like this have historically created as many new roles as they have disrupted. The transitions that went best were the ones where people could see what was coming clearly enough to prepare.”
Baumol's cost disease describes how industries with stagnant productivity face rising costs because wages must remain competitive with more productive sectors, causing prices in sectors like childcare, education, healthcare, and legal services to rise while prices in productive sectors like electronics fall.
“Economists call this Baumol's cost disease, and the clearest way to see it is to look at what happened to prices over the last 50 years. The things that got more productive, like electronics and computers, got cheaper, while the things that couldn't get more productive, like childcare, education, haircuts, healthcare, and legal services, kept getting more expensive.”
William Baumol and William Bowen observed in 1965 that a string quartet required the same number of musicians and time to perform Beethoven in the 19th century as in the 20th century, yet concert costs rose dramatically because musicians' wages had to rise to remain competitive with more productive industries.
“To understand the pattern, you need to go back to 1965, when a Princeton economist named William Baumol and his colleague William Bowen noticed something odd about the performing arts. A string quartet performing Beethoven in the 19th century required four musicians and about 25 minutes. A string quartet performing the same piece one century later required the exact same number of musicians and lasted about the same time. Nothing about the performance had gotten more efficient”
Approximately 30% of the US workforce has essentially zero AI exposure and works in physical, relational, hands-on occupations like cooking, mechanics, nursing, plumbing, bartending, and childcare that language models cannot replicate.
“about 30% of the workforce has essentially zero AI exposure. This includes cooks, mechanics, nurses, plumbers, bartenders, childcare workers, people doing physical, relational, hands-on work that no language model can replicate.”
For computer and math workers, AI is theoretically capable of handling around 94% of their tasks, but in observed professional use it is currently only doing about 33%, with similar gaps appearing in legal work, architecture, and engineering.
“For computer and math workers, AI is theoretically capable of handling around 94% of their tasks, but in observed professional use, it's currently doing about 33%. And a similar pattern shows up in legal work, in architecture and engineering, and across many other professions.”
Entry-level employment in AI-exposed occupations has already dropped 14% compared to the pre-ChatGPT era, while entry-level job postings across the US have dropped 35% since January 2023, suggesting the leading edge of disruption is already visible.
“Entry-level employment in AI-exposed occupations has already dropped 14% compared to the pre-ChatGPT era. And that's probably just the beginning because job postings fall before employment. And entry-level job postings across the US have dropped 35% since January 2023.”
Existing economic metrics like GDP, unemployment, and wage data were designed to count jobs and people but were never built to measure which tasks within jobs AI can perform, so disruption doesn't register in official numbers until well after it's underway.
“GDP, unemployment figures, and wage data were designed to count jobs and people, and they do that reasonably well. But, they were never built to look inside a job and ask which parts of it AI can already technically perform. So, the change doesn't show up where we're looking for it. By the time the disruption shows up in the official numbers, it's already well underway”
IBM replaced a chunk of its HR department with AI tools, while Salesforce stopped hiring engineers and lawyers because its CEO stated that AI can do the work.
“IBM replaced a chunk of its HR department with AI tools, while Salesforce stopped hiring engineers and lawyers because, as its CEO put it, AI can do the work.”
In 2025 alone, more than 100,000 job losses were linked to AI restructuring, and AI is already writing more code every day than every human developer on the planet combined.
“In 2025 alone, more than 100,000 job losses were linked to AI restructuring. AI is already writing more code every day than every human developer on the planet combined.”
If the Iceberg Index is not used, the cost is a two-speed economy: one half dramatically more productive, the other steadily less affordable, with fiscal pressure landing hardest on exact services societies can't function without.
“if they don't, the cost is a two-speed economy, one half dramatically more productive, the other steadily less affordable, with the fiscal pressure landing hardest on the exact services societies can't function without.”
The anxiety about AI in headlines has been aimed at roughly 1/5 of the actual problem, with the other 4/5 sitting on a fault line that no government, company, or individual worker has been preparing for.
“This means the anxiety about AI that's been dominating headlines for 2 years has been aimed at roughly 1/5 of the actual problem. The other 4/5 have been sitting on a fault line that no government, no company, and no individual worker has been preparing for.”
Most work that AI cannot touch is essential and people cannot stop using it when prices rise (healthcare, education, elder care, skilled trades), yet most of these services are either funded or subsidized by governments that are already stretched thin.
“People don't pull their kids out of school because costs went up or skip the plumber when the pipes burst. Healthcare, education, elder care, and skilled trades aren't things people can just stop using when prices rise. And most of them are either funded or subsidized by governments that are already stretched thin.”
The decision to weight the Iceberg Index by wage value rather than time spent is critical because automating 60% of someone's time does not mean automating 60% of their value if the remaining time involves complex judgment.
“An accountant might spend 60% of their time on document processing and data entry, and 40% on complex judgment calls and client relationships. Those are very different things, and treating them as equivalent would tell you almost nothing useful. Because automating 60% of someone's time doesn't mean automating 60% of their value. Weighting by wage value, on the other hand, means the index reflects where the actual economic exposure sits.”
Billions being spent on workforce preparation may be systematically aimed at the wrong places because the tools used for planning cannot see the actual distribution of AI risk.
“That means the billions being spent right now on workforce preparation may be systematically aimed at the wrong places entirely.”
The Baumol cost disease argument is a reasonable extrapolation from the Iceberg Index findings rather than a finding in the study itself, and outcomes depend on undecided factors like what firms will do with the capability, how fast governments respond, and which workers successfully retrain.
“The Baumol argument is a reasonable extrapolation from what the study found, not a finding in the study itself. Because the Iceberg Index tells you where the skill overlap sits right now, not what firms will do about it, how fast governments will respond, or which workers will successfully retrain. Those outcomes depend on decisions that haven't been made yet.”
The Iceberg Index deliberately does not account for physical robotics (purely digital and cognitive AI tools), does not measure outcomes or job losses (only technical capability), and does not predict when disruption will happen—it functions more like an earthquake risk map showing which buildings sit on fault lines rather than a weather forecast predicting if and when tremors will hit.
“There are a few things the Iceberg Index deliberately does not do, and it's worth being clear about them. It doesn't account for physical robotics. For now, this is purely about digital and cognitive AI tools. And it measures technical capability only, not outcomes, which means it doesn't predict job losses or forecast when any of this will actually happen. What companies do with this information, when they do it, and whether regulators allow it, none of that is in the model. Think of it less like a weather forecast and more like an earthquake risk map.”
It's like trying to navigate a new city using a 30-year-old map—the streets look familiar, but nothing is quite where you'd expect.
“It's a bit like trying to navigate a new city using a 30-year-old map. The streets look familiar, but nothing is quite where you'd expect.”
The friction between technical AI capability and actual deployment is driven by regulation, integration challenges, and the requirement that organizations maintain human oversight, but these friction points may resolve as technology matures.
“The technical capability is already there, but right now it's being held back by regulation, integration challenges, and the simple fact that most organizations still require a human to check AI's work. These are all friction points that may resolve themselves as technology matures.”
Ohio and Michigan follow the same pattern as Tennessee: they have spent years preparing for robots to take over factory floors, but the white-collar disruption from AI is arriving first.
“Ohio and Michigan follow the same pattern. These states have spent years preparing for robots to take over the factory floors, but the white-collar disruption is arriving first.”
The map most people are using for workforce planning was drawn for a different economy, and the Iceberg Index represents an attempt to draw a better one.
“The map most people are using was drawn for a different economy. The Iceberg Index is an attempt to draw a better one.”
In the industrial era, output per hour had to be invented because nobody had a way to capture physical productivity at scale, establishing the historical precedent that new economic eras require new measurement frameworks.
“In the industrial era, output per hour had to be invented because nobody had a way to capture physical productivity at scale.”