AI Impact Tracker
Sourced data. Real case studies. The questions that matter.
AI Labor Impact Scorecard
Key metrics framing the AI-labor transformation. Sources cited — not predictions, not opinions.
What we're tracking — the unknowns where certainty doesn't exist yet.
Will AI job creation outpace displacement?
Historical precedent says yes — ATMs didn't eliminate bank tellers, and spreadsheets created more accounting jobs. But the pace and breadth of generative AI is unprecedented.
How fast is displacement happening?
Klarna cut 57% of its workforce in 3 years. Is that the template, or an outlier? Most companies are augmenting, not replacing — so far.
Does AI compress wages even if job counts stabilize?
If AI makes every worker 3× more productive, companies need fewer workers. Even employed workers may face downward wage pressure as supply of capable labor effectively expands.
Which sectors are most exposed?
Administrative, legal, and financial services show highest exposure. Physical trades and healthcare show lowest. But AI is advancing into coding, design, and creative work faster than predicted.
Labor Market Data
Live FRED data — the numbers behind the narrative.
Indexed to 100 at Jan 2000 — divergence shows whether employment keeps pace with economic output.Source: FRED (PAYEMS, GDPC1)
- Nonfarm Payrolls
- Real GDP
The portion of economic output going to workers vs capital. Declining share = more value captured by technology/owners.Source: FRED (PRS85006173)
The productivity-pay gap: output per hour grows faster than what workers take home. The widening gap is where AI's economic impact concentrates.Source: FRED (OPHNFB, COMPRNFB)
- Productivity (Output/Hr)
- Real Compensation
The conversation around AI and jobs has been dominated by two extremes: utopian visions of human liberation from tedious work, and dystopian fears of mass unemployment. The reality, as with most technological transitions, will likely land somewhere in between — but with characteristics that make this transition genuinely different from past ones.
What's different this time: Previous automation waves targeted routine physical tasks — assembly lines, agriculture. AI targets cognitive tasks, including analysis, writing, coding, and decision-making. This puts white-collar, knowledge-worker jobs in the crosshairs for the first time at scale.
Goldman Sachs estimates that 300 million jobs globally could have significant portions of their tasks automated by generative AI. This doesn't mean 300 million people lose their jobs — it means 300 million roles will be fundamentally restructured. Some tasks within those roles will be automated; others will be augmented.
The World Economic Forum projects 97 million new roles emerging in AI-adjacent fields — AI training, prompt engineering, AI ethics, data labeling, and new categories we haven't named yet. Whether creation outpaces displacement depends on the speed of transition and the adaptability of the workforce.
Greg Ip, WSJ Chief Economics Commentator
"Labor's share of GDP has fallen 7 percentage points over 40 years. AI could accelerate that trend — or, if it creates enough new demand, reverse it. The key variable isn't the technology. It's the policy response."
Sector Exposure
Not all sectors face equal AI exposure. Administrative and clerical roles show the highest displacement risk (46% of tasks automatable), followed by legal services (44%) and financial operations (43%). Meanwhile, physical trades (construction, plumbing, electrical) show less than 6% task exposure, and direct healthcare shows under 10%.
The most interesting category is augmentation rather than replacement. Software engineering, for example, shows 30-40% task automation potential — but early adopters report that AI tools make them more productive, not unemployable. GitHub reports that developers using Copilot complete tasks 55% faster. The question is whether companies use this to do more work with the same staff, or the same work with less staff.
Historical Context
The ATM was supposed to kill bank teller jobs. Instead, it lowered the cost of running a branch, banks opened more branches, and teller employment actually grew for 20 years after ATM introduction. Spreadsheet software was supposed to eliminate accountants — instead, it made financial analysis accessible to every business, and accounting employment expanded dramatically.
The counter-argument: those transitions played out over decades. AI adoption is happening in months. Klarna deployed AI customer service in 2023 and reduced its workforce from 5,000 to 3,800 within two years. The company's revenue per employee jumped from $300,000 to $1.3 million. That's a compression of the displacement-to-adaptation cycle that has no historical precedent.