Six updated indicators, a sharper look at responsible AI, and what it means for your scores

How do you measure returns on AI? A recent piece in  The Economist  makes the case that usage numbers alone say very little, and that speed without quality isn’t a return at all. For investors, that raises a second question: how are the companies in your portfolio governing the AI they deploy? 

It’s a question our October 2026 quarterly release takes on directly. 

Responsible AI, measured more broadly

Our Responsible AI Policy indicator previously recognized only companies that were members of EqualAI. Starting this quarter, it measures AI governance frameworks, responsible AI policies, or training to mitigate AI bias. This gives a more complete picture of how companies are addressing AI risk internally.   

We also continue to track AI-related harm. For companies in Communication Services and Information Technology, our Adverse Effects indicator counts substantiated cases of AI, facial recognition, or surveillance products used in discriminatory ways. 

Both of these advances are the beginning steps of a more comprehensive buildout at As You Know on measuring the impacts, positive and negative, of the watershed social and economic change occurring through largely unregulated AI development and implementation.  Stay tuned for more to come from us on tracking AI value destruction and creation. 

AI’s physical footprint

Responsible AI isn’t only about algorithms. It’s also about the data centers that power them. As tech companies race to build, investors are asking how that expansion fits with their climate commitments, and which communities bear the impact of where data centers are built. We’re watching this space closely. 

SPOTLIGHT: As You Sow on AI data centers