The simpler first phase of AI adoption—handing people a chatbot—is giving way to a harder one: rewiring the systems, data and workflows underneath.

While helping people cope with burnout matters, it cannot compensate indefinitely for unhealthy work conditions and organizational norms.

While using AI to take over decision-making when it comes to termination may not be a solid idea, using it for assistive work here is more defensible.

Without clear expectations about human oversight, employees using AI could drive up organizational risk, finds a new report.