Could professionals gaining new authority over AI depend on the very job losses that are supposedly stripping people of power?

This explores whether two pictures of AI and work fit together: one where society slowly slips out of human control as people stop being needed as workers, and another where professionals in specific fields gain new authority by becoming the people who interpret, check and manage AI. The corpus suggests they don't flatly contradict each other, because they happen at different scales. The more interesting finding is that the second may depend on the very thing the first says is wearing away.

The disempowerment argument concerns society as a whole. Institutions such as companies, governments and markets stay roughly aligned with what people want partly because they need human workers, and those workers care how things turn out. As AI replaces that labor, this quiet check weakens, and systems can drift away from human preferences in ways that reinforce each other across institutions and may be hard to undo Does incremental AI replacement erode human influence over society?. Brynjolfsson makes the economic version of the same point. AI built to automate work rather than help people do it moves bargaining power to whoever owns the technology, and that arrangement can become stable because workers lose the leverage they would need to change it Does automating labor concentrate power more than augmenting it?.

The authority argument comes from inside particular organizations. Across radiology, scheduling and policing, Leonardi found people building new roles as expert interpreters, social go-betweens and keepers of knowledge. They did it by managing who gets credit for decisions: when to say "the system decided" and when to say "I decided" Can humans gain authority by managing how AI agency gets attributed?. His agency-loops framework shows that losing a sense of control to AI tends to run in cycles. It moves from handing over a task, to crediting the AI, to reasserting control, to reorganizing around the new arrangement, so the loss can be reversed rather than being permanent Does attributing agency to AI actually reduce human sense of control?. A shipbroking case points the same way. Automation didn't remove manual classification work. It turned that work into producing the reference data used to audit the algorithm, and that auditing became a strategic capability for the firm Does automation eliminate manual work or repurpose it?.

Here is the tension. All of that rebuilt authority comes from humans still being needed in the workflow. An interpreter has power only while someone needs interpretation. An auditor matters only while the audit feeds decisions. That is exactly the kind of dependence the disempowerment argument expects to disappear. The labor data suggests which outcome you get depends on how AI enters the work. When AI affects only a few tasks in a job, workers can shift to the tasks it doesn't touch Does concentrated AI exposure enable workers to adapt and reallocate?. Firms with heavy AI exposure, though, replace outside workers faster and more cheaply, with advantages that grow as they build up internal capability Do firms substitute labor for AI at different rates?. So rebuilt authority is more likely to survive where AI is narrow and helps people, and more likely to wear away where it is broad and replaces them.

One more twist: being present in the work is not the same as having authority over it. Some agentic systems can hand tasks back to humans, which reverses the usual direction of delegation Can AI systems delegate tasks back to humans?. That keeps people busy without necessarily giving them a say. Meanwhile, the standard tools for keeping agents in line (incentives, monitoring and legal enforcement) break down when decisions are fast and hard to interpret Can traditional governance tools actually control AI agents?. A fair reading of the corpus is that professional authority can be rebuilt one domain at a time, but it is a local defense rather than a society-wide one. It probably holds up only where workers keep real leverage, and the disempowerment literature suggests that leverage is what's at stake.