Nobody knows how to measure AI's economic impact, but this paper tries to fix that
What happened
This paper reviews how companies use AI and how it affects their business. It finds that different ways of measuring AI lead to different conclusions about its impact. The paper then offers a way to choose the right measurement for different situations.
Why it matters
Companies and governments are pouring money into AI, but they often don't know if it's actually working or what it's changing. This paper is a guide for figuring out what to measure and how. It means that for the first time, people trying to understand AI's real-world effects have a common language and a set of tools to use.
The signal
Watch for other researchers and economists to adopt this paper's framework for measuring AI's impact, or for new datasets to emerge that align with its recommendations.