Two of the nation's leading law schools recently responded to the same technological development in strikingly different ways. The University of California, Berkeley School of Law adopted one of the country's most restrictive classroom AI policies. Earlier this year, the law school announced a policy prohibiting students from using AI to conceptualize, outline, draft, revise, translate, or edit work submitted for credit. Students may use AI only to identify potential legal sources for research while remaining responsible for every other aspect of their work. The University of Chicago Law School has largely taken the opposite approach, leaving decisions about AI primarily to faculty judgment.
The obvious question is which institution has the better AI policy. The more interesting question is why two highly respected institutions, confronting the same technology, would reach such different conclusions. The longer I considered that question, the less I believed the answer had much to do with artificial intelligence itself. I found myself becoming less interested in which institution was "getting AI right" than in what each policy revealed about how institutions allocate human capacity when technology changes the economics of work. Viewed that way, the policies began to look less like isolated responses to AI and more like expressions of institutional priorities.
Artificial intelligence changes the economics of producing knowledge work. Research that once required hours may now take minutes. Drafts can be assembled more quickly. Routine administrative work increasingly requires less time than it once did. Whether institutions welcome those changes or resist them, they all encounter the same reality. When the economics of work change, human capacity changes with it.
What happens after that point is no longer determined by the technology. It is determined by the institution. Leaders decide whether newly available capacity will be 
To make those decisions visible, I architected the Human Capacity Allocation Framework. The framework shifts the unit of analysis from artificial intelligence itself to the institutional decisions governing where human capacity should remain central, where it should be redistributed, where it should be developed, and where automation is appropriate.
It does not assume that every gain in efficiency should lead to the same institutional response. Capacity may be preserved when the work itself is developmental, redistributed toward work requiring greater human judgment, developed where people need new capabilities, or released to automation when the work no longer requires sustained human attention. The framework therefore does not ask whether an institution has adopted the "right" AI policy. It asks: What allocation decision is the institution making?
Viewed through this lens, the contrast between Berkeley and the University of Chicago reveals more than two positions on artificial intelligence. Berkeley allocates human capacity toward preservation. Its policy protects the developmental process through which students learn to conceptualize, draft, revise, and ultimately form legal reasoning. Chicago allocates capacity toward redistribution. By leaving greater discretion to faculty, it creates room to move student and faculty attention toward learning how to practice, judge, and teach alongside AI. These are not simply different AI policies. They are different judgments about what legal education exists to produce and where human capacity should remain concentrated.
