An essay discussing the impact of AI on knowledge work has gained 84 million views on X. Observations indicate that significant portions of the content were generated by AI, raising concerns about the quality and authenticity of writing.

Semantic ablation refers to the algorithmic erosion of high-entropy information in AI writing, leading to generic and boring outputs. This phenomenon is a structural byproduct of greedy decoding and reinforcement learning from human feedback (RLHF).

The study investigates the cognitive effects of using large language models (LLMs) for essay writing by comparing performance and neural activity among participants using LLMs, search engines, and no tools. A total of 54 participants completed three sessions, with 18 continuing to a fourth session that involved switching between LLM and brain-only conditions.