The great value migration

A useful preprint (Future of Work with AI Agents: Auditing Automation and Augmentation Potential across the U.S. Workforce; link) by Stanford University's Yijia Shao, Humishka Zope , Yucheng Jiang, Jiaxin Pei, David Nguyen, Erik Brynjolfsson, Yang Diyi provides a framework and some data to think about the sensitivity of many occupations to AI advances. The authors introduce the "Human Agency Scale (HAS) as a shared language to quantify the preferred level of human involvement" in an occupation. They find a spectrum of HAS profiles across occupations and offer this as an audit tool to assess the potential for displacement. However imperfect, I find their framework actionable and practical 💡.

Of particular interest to me are the results summarized in their figure 7 (Comparing skill rankings by average wage and required human agency). I have been predicting for 25 years that we would witness a migration of value toward individual skills that involve (1) judgment and/or (2) empathy and interpersonal relationship mastery. I may have been a bit early (I thought it would happen by 2010 for sure; that's why you need to take punditry with a boulder of salt) but this shift is happening now and it is very clear in the HAS framework: on the figure, "skills are ranked by average wage (left) and average required human agency (right). The figure highlights the top five skills with the largest upward (green) and downward (red) shifts in rank, suggesting a potential shift in valued workplace skills—from information processing toward interpersonal and organizational competencies."

I think this left-brain-to-right-brain value migration is telling only part of the story: the ability to combine and exercise both sides of the brain will become the clear nexus of human value, for example the ability to analyze data AND apply personal experience to define and prioritize actions. That is why the ability to think computationally or to understand the outputs of AI remain essential skills: even of you don't ever code again after college, you know how to evaluate code-related products.

The key point to keep in mind is that many occupations of the future (say, the next 10 years) are not on that list because they don't exist!

(@chloe jian ma, San Francisco)