The Challenge

The advent of Transformative AI¹, or TAI, would yield economic change on the scale of the Industrial Revolution but has the potential to occur radically faster. TAI could boost productivity levels and accelerate scientific progress, putting us in a realm beyond traditional economic modeling. At the same time, TAI risks disruption in labor markets and changes in the concentration of wealth and power. With foresight, we can adapt our institutions and create a future of widely shared prosperity.

The gap between rapidly growing technological capabilities and slowly improving economic understanding, skills, institutions, and policies is the crux of the coming decade’s societal challenges. Economics must be transformed in the face of such technological change.

There is an urgent need to understand the economic implications of Transformative AI: If machines are created that can outperform humans on a significant share of tasks, many of our existing institutions, norms, and systems will need to be reinvented. This raises questions regarding distribution, concentration, inequality, information flows, geopolitics and trade, AI safety and alignment, well-being, and a host of other important topics.

The Digital Economy Lab is actively engaged in developing a robust economic research agenda to address the potential challenges associated with TAI. The papers linked below represent a subset of ideas and thinkers we believe are shaping this conversation (email sdel-admin@stanford.edu if you have a suggested addition).

¹ For our purposes, a working definition of TAI is provided by Karnofsky, 2016. “Transformative AI refers to potential future AI that precipitates a transition comparable to (or more significant than) the agricultural or industrial revolution. The concept of “transformative AI” has some overlap with concepts put forth by others, such as ‘superintelligence’ and ‘artificial general intelligence.’ However, ‘transformative AI’ is intended to be a more inclusive term, leaving open the possibility of AI systems that are ‘transformative’ despite lacking many abilities humans have.”