---
id: "lib-2026-07-13-we-must-act-now"
type: "library"
title: "We Must Act Now"
lang: "EN"
org: "Brynjolfsson Agrawal Korinek Cunningham"
source_type: "Article"
date: "2026-07-13"
week: "2026-W29"
url: "https://nosov.ai/library/lib-2026-07-13-we-must-act-now/"
original_url: "https://www.wemustactnow.ai/"
tier: "S"
tier_name: "Shapes the field"
tier_reason: "A one-paragraph statement whose weight comes from its signatories: 17 Nobel laureates in economics including Acemoglu, Stiglitz, Krugman and Bernanke, plus Brynjolfsson, Autor, Bengio and senior staff from Anthropic, OpenAI and Google."
summary_basis: "full-text"
original_minutes: "3"
tags: ["policy", "complementarity"]
---

# We Must Act Now

Tier S (Shapes the field): A one-paragraph statement whose weight comes from its signatories: 17 Nobel laureates in economics including Acemoglu, Stiglitz, Krugman and Bernanke, plus Brynjolfsson, Autor, Bengio and senior staff from Anthropic, OpenAI and Google.

Summary based on the full text.

## In short

"We Must Act Now: A Statement on AI's Transformation of the Economy" is a short public statement hosted by Stanford, first listed by Erik Brynjolfsson (Stanford), Ajay Agrawal (Toronto) and Anton Korinek (Virginia and Anthropic). It says AI may become radically more powerful within ten years and could drive an economic shift larger than the Industrial Revolution over a much shorter period. It asks economists, policymakers and technology leaders to start now on the economics of transformative AI and on the incentives and institutions that steer it toward complementing people. The document contains no data; its signal is who signed it.

## Key numbers

- 17: signatories marked as Nobel laureates on the published list
- 10 years: horizon over which the statement says AI may become radically more powerful

## What it reports

- The full text is one paragraph of about 100 words, followed by a signature list of more than 200 names.
- It names two outcomes side by side: risks "including large-scale job displacement" and opportunities "such as major gains in living standards".
- 17 signatories are marked as Nobel laureates, among them Daron Acemoglu, Joseph Stiglitz, Paul Krugman, Ben Bernanke, Michael Spence, Christopher Pissarides, Philippe Aghion, Peter Howitt, Robert Shiller and George Akerlof.
- Labour and productivity economists who rarely sign joint statements appear together: David Autor, Lawrence Katz, Nicholas Bloom, John Van Reenen, Raffaella Sadun, Chad Jones, Jason Furman, Olivier Blanchard and Gita Gopinath.
- Lab and industry signatories include Jack Clark (Anthropic), Jeff Dean (Google), Wojciech Zaremba and Noam Brown (OpenAI), Yann LeCun, Yoshua Bengio and Eric Schmidt.
- Known sceptics of AI hype, such as Arvind Narayanan and Daron Acemoglu, sign alongside people who expect fast progress, such as Daniel Kokotajlo.
- The statement proposes no specific policy, budget or timetable.

## How they got it

This is a position statement, not a study. It was published as a web page with an open signature list and a Stanford contact address for anyone whose name was added in error.

## Limits

- No evidence, model or forecast is offered; "may become radically more powerful" is a conditional claim.
- A number of signatories work for frontier labs or invest in AI companies, which gives them a stake in the framing.
- Signing a one-paragraph text with broad wording costs little and does not mean each signatory agrees on timing or magnitude.

## Why it matters for companies

When economists who disagree on most AI questions sign the same sentence about possible large-scale job displacement within a decade, the scenario has moved from fringe to mainstream planning territory. Workforce plans, reskilling budgets and organisation design built on a five-to-ten-year horizon should now carry an explicit high-change scenario next to the base case. The statement gives no guidance on what to do, so its use inside a company is as a reason to run that scenario, not as evidence for any particular outcome.

Original: https://www.wemustactnow.ai/