Allow me to get to know you, mistakes and all
Growing use of large language models like ChatGPT and Claude to write emails, Slack messages, and performance reviews is prompting concern that everyday communication is becoming generic, bloated, and less authentic. Many argue that AI-written text erases individual voice, introduces an effort imbalance between writer and reader, and can undermine trust, especially in personal or high-stakes contexts. Others counter that these tools are indispensable for non-native speakers, people with disabilities, or those facing hostile corporate cultures, and see AI as a legitimate aid for clarity, productivity, and overcoming writing blocks when used thoughtfully.
Frustration with AI-Generated Workplace Communication
- Many dislike obviously-LLM-written Slack, email, GitHub issues, and PR descriptions; long, polished paragraphs are now often a negative quality signal.
- Complaints focus on verbosity, buzzword padding, low “signal-to-token” ratio, and the sense of reading hollow “AI slop.”
- Using AI for critical 1:1 feedback (e.g., performance reviews) is seen as especially jarring and dehumanizing.
Authenticity, “Voice,” and Social Expectations
- One side: AI-polished text robs others of seeing real quirks, mistakes, and thought patterns; it flattens personality and makes everyone sound the same.
- Counterpoint: No one is entitled to another’s “authentic self”; people routinely curate their public face, and using tools (books, coaches, LLMs) is just another form of that.
- Disagreement over whether colleagues have a legitimate interest in your “real voice” or only in clear, functional communication.
Efficiency, Risk Management, and Asymmetry of Effort
- Some workplaces discourage ChatGPT/Claude for internal comms as unproductive and alienating; basic spell/grammar tools are accepted.
- Others rely heavily on LLMs to handle large volumes of repetitive questions, drafts, and documentation, claiming big productivity gains.
- Several note an effort asymmetry: “I couldn’t be bothered to write it, but you have to read it,” which is perceived as disrespectful.
Non-Native Speakers, Disabilities, and Accessibility
- Non-native English speakers and some disabled contributors say LLMs are a crucial equalizer for clarity and credibility.
- Others respond that minor grammatical errors are fine; sloppiness is the problem, not imperfect English.
- Some fear polished AI text is now less trusted than imperfect but clearly human language.
AI as Writing Tool vs Thinking Tool
- Distinction drawn between:
- AI as output tool: generating or heavily rewriting messages, which often erases personal style.
- AI as thinking tool: rubber-ducking, structuring ideas, overcoming blank-page anxiety, then writing/editing in one’s own words.
- ADHD and “blank page” users describe AI as a powerful starter, but others warn this may atrophy core planning and drafting skills.
Language Flattening and Cultural Effects
- Multiple comments describe AI as a “smoothing function” or “genericizer” that homogenizes style and vocabulary.
- Some claim early evidence that mainstream language is shifting toward AI patterns (e.g., more em dashes, certain stock phrases).
- Fears that pervasive AI-written text will reshape human writing norms, making everything more generic—while also pushing some people to become more idiosyncratic to stand out.
Norms, Labels, and Future Use
- Proposals include standardized “human-only” labels for content and clearer norms about when AI use is acceptable (e.g., grammar vs full generation).
- Others argue it’s too early to draw hard lines; society is still experimenting, and future uses (personal PR, automated coordination, richer relationships) are uncertain.