Today, the Stanford @DigEconLab launches the AI Economic Indicators, a new platform for tracking how AI is reshaping work, productivity, adoption, and the economy.

1/6

We need more real time data on how AI may be impacting the economy - this is a really useful addition.

Each month, we will update it with new data, and periodically, we'll add new metrics and dashboards.

Check out the website: https://digitaleconomy.stanford.edu/project/indicators/

And let me know what you think!

6/6

Three key components:

1. The Canaries Dashboard (jointly with ADP Research) — labor market outcomes across occupations and worker groups with different levels of AI exposure.

This is the same data that we used for our "Canaries in the Coal Mine?" paper, updated monthly.

2/6

2. The Takeoff Tracker — macroeconomic indicators to give us an early warning of when and where better AI capabilities lead to an economic take-off.

3/6

The site (https://indicators.stanford.edu/) has three components at launch: (1) the Canaries Dashboard, (2) the Takeoff Tracker, and (3) the Adoption Monitor. Each will be updated regularly---(1) and (2) monthly.

3. The Adoption Monitor — AI adoption by consumers, workers and firms across multiple datasets.

4/6

It's been a thrill to turn this idea into a public resource with colleagues across the Lab. Special thanks to @connacher_ , who leads the project, along with @therealcko, Susan Young, and Matty Smith — and to the many researchers behind it, including @NelaRichardson, @BharatKChandar, @RuyuChen, @Andrew Wang, @pawtrammell, and @I_Am_NickBloom.

5/6

(3) summarizes adoption survey work from many great researchers — all credit to them! Includes @I_Am_NickBloom, @iyotzov, @ProfDavidDeming, @Jon_Hartley_, @FilipJole, @BrendanDMoore, and many researchers I couldn't find on Twitter.

First, individuals.

The Canaries Dashboard, a collaboration between @AdpResearch and the Lab, continues the work of our Canaries in the Coal Mine study as it tracks employment trends at varying levels of AI exposure. http://canaries.stanford.edu

In (2), @pawtrammell, along with @erikbryn and me, updates and extends Nordhaus (2021), "Are We Approaching an Economic Singularity?" His answer: "no." Our answer: "still no." But we will be updating the results monthly. https://www.aeaweb.org/articles?id=10.1257/mac.20170105

May data coming soon!

We're also now integrating the job transitions framework data from @Alex_M_Richmond, @RonnieChatterji, and co.

The original paper details robustness checks.

I plan to write more on how Canaries fits into the portfolio of AI x labor market measurement

We have plans to expand each of these three products, including new cuts of Canaries (e.g., by gender) as well as next-generation exposure measures, new indicators of takeoff (e.g., unmeasured consumer surplus in GDP), and frontier model usage data.

The AI Economic Indicators is live! Our new, free platform tracks how AI is reshaping jobs, productivity, and the economy.

Three unique dashboards at launch for you to explore, with more on the way: http://indicators.stanford.edu

@BasilHalperin I've interest-rate pilled (now realizing I should've credited you, Chow, and Mazlish in the tweet below)

And we hope to add more pages on changing skill demands, inference market structure, and model capabilities. Lastly, we look forward to expanding our geographic scope beyond the U.S.

Please get in touch with questions, comments, or ideas for extensions!

The Takeoff Tracker looks for signs of explosive growth and AI-based economic takeoff, evaluating indicators by strength of evidence. http://takeofftracker.stanford.edu

@connacher_ @DigEconLab @erikbryn been hoping for something just like this forever

@emollick What do you make of recent economics-focused benchmarks like OpenAI’s GDPval and Andon’s Vending Bench?

Seem directionally closer for measuring real-world impact vs traditional evals

@erikbryn @DigEconLab Really impressive work

Any collabs coming with OpenAI’s GDPval or Andon’s Vending-Bench?