Special ── 2026-10-23
Stanford HAI Launches Monthly AI Labor Indicators
Stanford HAI's monthly AI indicators track early-career job declines in real time
What changes
According to Stanford HAI, the effects of AI on employment can now be checked with monthly updates. Official statistics report longer-term trends. The new tools capture swift shifts as they happen.
Managers and HR leaders can use the indicators to test how exposed their own occupations are to AI. Stanford HAI reports that employment has dropped in software development and customer service.
Why it matters
Stanford HAI's Connacher Murphy says AI has been adopted perhaps more rapidly than any other technology ever. That makes it hard to separate genuine economic change from hype.
Bharat Chandar, the Lab's research director for AI economics and policy, says the team does not want to fly blind into a potentially very large economic transformation. He says definitive answers have not arrived, but the team is better placed to catch early signs.
What the report says
In June 2026, Stanford HAI's Digital Economy Lab (DEL) launched a set of dashboards called AI Economic Indicators. They are updated monthly and track AI's economic footprint in real time.
Stanford HAI reports that employment in AI-exposed jobs such as software development and customer service has dropped by more than 4% this year. Occupations that rely on AI automation are showing slowing or shrinking job growth.
According to the Canaries Dashboard, employment for 22- to 25-year-olds in those two sectors has dropped over 3% annually since ChatGPT's release in November 2022. In recent months the decline sharpened to 4.4%. Job growth for young women is slower than for young men, but Chandar cautions against reading too much into this because the ADP Research data may not be representative enough.
Using the Anthropic Economic Index, Murphy says occupations that skew toward an automation-heavy pattern have slower or contracting employment growth, especially among early-career workers. The Transformation Tracker, which follows 12 macroeconomic indicators, does not point to explosive growth so far.
Who and what scope
The dashboards come from the Digital Economy Lab at Stanford HAI. HAI is an institute that advances AI research, education, policy, and practice.
They cover employment by occupation and the uptake of generative AI among individuals and companies. They use data from private companies, academic researchers, and government statistical agencies. The Canaries Dashboard was built with ADP Research, a New Jersey-based institute. The article does not name the country or countries covered.
Why this report
Murphy says it is very hard to get a big, aggregate view of what is actually happening, what is hype, and what is genuine economic change. The indicators are meant to answer that problem.
Official statistics report longer-term trends, so fast shifts are hard to see. DEL built these dashboards to give a clearer and earlier picture of what AI is doing to the economy.
How they studied it
The project pulls data from private companies, academic researchers, and government statistical agencies. It organizes the data into three primary dashboards.
The Transformation Tracker monitors 12 macroeconomic indicators. The Canaries Dashboard uses anonymized payroll data from ADP Research's parent company and groups workers by high and low AI exposure. The Adoption Monitor relies on surveys to track generative AI uptake among individuals and companies.
For automation versus augmentation, the researchers use the Anthropic Economic Index. It analyzes roughly a million anonymized conversations with the model Claude, tagged by user occupation. Automation is when technology such as AI performs specific tasks according to instructions defined by a human.
The new GDP-B: Surplus Observer collects consumer surplus data through online surveys, led by David Nguyen. Consumer surplus is the difference between what someone is willing to pay and the actual price. The goods covered include generative AI tools, coffee, online search, and GLP-1 drugs. Stanford HAI does not give survey sample sizes or the size of the ADP data.
Key points
- In June 2026, Stanford HAI's DEL launched three monthly-updated AI Economic Indicators dashboards.
- Employment for 22- to 25-year-olds in software development and customer service fell over 3% annually, and 4.4% recently.
- The Transformation Tracker's 12 macro indicators do not currently point to explosive growth.
Conclusion
Stanford HAI's indicators offer a monthly way to check how AI is affecting employment. They show employment declines in AI-exposed occupations for early-career workers.
However, Chandar does not treat findings such as the gender gap as settled. Definitive answers have not arrived, so later updates will matter.
Quotes and sources
“Employment in AI-exposed jobs like software development and customer service has dropped by more than 4% this year, while occupations that are reliant on AI automation are showing slowing or shrinking job growth.”── Stanford HAI, 2026-10-22
“What we find is occupations that really skew toward an automation-heavy pattern have slower employment growth or even contracting employment growth”── Stanford HAI, 2026-10-22
“We don’t want to fly blind into what could potentially be a very large economic transformation”── Stanford HAI, 2026-10-22
Source articles
- Stanford HAI 22 Oct 2026
Written by AI (Claude) from the report and its coverage, then checked by Jev (TypeSafe's judgement model), a Gemini fact-check and a verbatim check of every quote.