Missing rules for working with AI don't hit everyone equally — so who actually pays the price when things go wrong?

This explores whether missing rules for human-AI work land harder on frontline staff than on leaders. The corpus pushes back on the premise first. BCG's survey of nearly 12,000 employees found that agent deployment more than doubled while half of respondents had no clear governance for human-AI teams, but worries about accountability were roughly equal at every level of the organization How far behind is AI governance compared to agent deployment?. So the gap itself isn't felt more at the bottom. What differs is who has to absorb the consequences and who has the power to close it.

The clearest evidence is that a worker's success with AI depends mostly on things they don't control. Microsoft's survey of 20,000 AI users attributes about two-thirds of AI impact to organizational factors such as culture, manager support and incentives, and only about a third to individual effort. Just 26% report clearly aligned leadership, and only 13% are rewarded for reinventing how they work Why do ready workers struggle to transform their work?. BCG reaches a similar conclusion from another angle: clear strategic direction predicts value better than access to tools Does strategic direction matter more than AI tool access?. A frontline employee with good tools and no direction is ready to change but has no permission or reward for doing it.

The gap also shows up in expectations and in anxiety. Executives in an NBER survey of nearly 6,000 firms predict AI will cut jobs, while employees at the same firms expect it to add them Do executives and employees agree on AI's job impact?. In Capgemini's survey, 61% of leaders report rising employee anxiety about agents, yet fewer organizations are prioritizing reskilling Why aren't organizations investing in reskilling despite AI job displacement fears?. Leaders are planning around one future while employees prepare for another, and nobody is paying for the move between them.

The less obvious finding is that frontline workers often end up as the governance layer by default. In a field experiment on Alibaba's Taobao, human agents who took over after AI failures fixed technical problems but couldn't recover customers who were already frustrated, and late handoffs failed When human workers take over from AI, does it actually fix the problem?. When nobody has defined when a human should step in, the worker inherits the damage. Managers short on attention also hand decisions to machines even when human judgment would be better Do managers delegate to machines even when human judgment is better?. Systems that look competent spread accountability across so many people and components that no one clearly owns a failure How do competent systems quietly undermine safety oversight?. In practice, the person closest to the output is left holding the blame.

There's also a skills angle. Under mandatory use, GPT-4 helped lower-performing consultants most, which narrowed the gap with top performers, but the effect may reverse when adoption is optional Does AI help lower-skilled workers catch up to top performers?. Clunky chatbot interfaces also hurt less experienced users most Is the AI capability gap really an interface problem?. Missing rules about who must use AI, and how, can quietly decide whether it narrows or widens inequality inside a company. The corpus doesn't directly measure governance harm by job level, so this reading is pieced together from adjacent evidence rather than demonstrated outright.