Please don't use an LLM to communicate with other human beings
The article argues against using large language models (LLMs) to craft personal communications like emails or messages for other people, contending that doing so undermines authentic human connection and outsources the personal effort required for genuine relationships.
Background
- LLMs (Large Language Models) like ChatGPT and Claude have become popular tools for writing emails, messages, and other workplace communication since late 2022. Many people use them to draft or polish messages to colleagues, clients, or even friends and family.
- The author argues that using an LLM for personal communication shifts the burden of understanding from the writer (who should work to be clear) to the reader (who must decode generic, bloated AI text), and that it cheapens relationships by outsourcing a fundamentally human act.
- Key distinction: using LLMs for tasks like summarizing, translating, or generating boilerplate is different from using them for one-on-one interpersonal communication where tone, nuance, and authenticity matter.
- The essay fits into a growing backlash against AI-generated communication, where people find it impersonal, dishonest, or inefficient — echoing concerns about "AI slop" flooding professional and personal exchanges.