Recent developments in computing infrastructure and natural language processing have unlocked previously unavailable capabilities in text processing, analysis, and generation. Their widespread use has resulted in calls whether it is possible to formally determine and detect LLM-generated output. This essay shortly presents the practical reasons for establishing detection methodologies for LLM-generated output and evaluates the technical feasibility of the task.
Access the Essay via: