Combining the various "tells" of an LLM (em dashes, delve, grammatical signs etc) with the context (Reddit comments vs professional setting), you could establis

Combining the various "tells" of an LLM (em dashes, delve, grammatical signs etc) with the context (Reddit comments vs professional setting), you could establish a rough probability it was AI generated. At this point, it's the best we can hope for.