IIR one of a NeurIPS papers (or similar) was a Walmart Labs paper on "Purpose driven AI Guide for Employees" or similar. Multiple mentions of psychological health in the same paper that describes an AI assistant that is hypothetically required for every employee of a certain stripe, that does give instructions and answer questions.
As a skeptic, the one-sided description of mental health, and the obvious parallel to a "always-on slave collar" jumped out. Is there legitimate use for this kind of employment technology?
As is so common, those most incentivized to produce and deploy such a thing, are the exact character that is not to be trusted with such an invasive machine. It just screams of a need to regulate, to my western culture eye.