that Microsoft AI study
While it is likely true that the 'reviewing' of LLM-generated output involves less critical analysis than doing the work oneself (for reasons that have been well-understood since before LLMs existed) - I don't actually think that a study based on self-reports with no external assessment is a credible source to back that up.
"How does one perceive their critical thinking" is just as prone to cognitive biases as the 'uncritical review' problem itself is.
Follow
that Microsoft AI study
(Or in other words: just because you agree with the conclusion, that doesn't mean that a given piece of research is sound, or well-executed, or even appropriate to the point being argued)
