By JuanDa Vasconez
The Silent Erosion: How AI is Changing Human Writing Skills
Something unsettling is happening to human expression. The gradual homogenization of language. Emails that echo one another. Reports with a polished but hollow tone. Creative projects that feel technically competent but mysteriously soulless. This erosion of individuality in writing is cause for concern. We are witnessing what researchers call the "flattening" of language (Hohenstein et al., 2023), a phenomenon where AI writing tools, despite their remarkable capabilities, are silently reshaping how humans communicate and, more worryingly, how we think about writing itself. Recent research provides concrete evidence of this effect: a cross-cultural study found that AI suggestions homogenize writing toward Western styles, with non-Western users altering their natural expression patterns to align with Western norms (Basu et al., 2023). However, this is not happening in a vacuum. It is part of a broader pattern of extraction that defines our economic system: the systematic harvesting of human capabilities, creativity, and labor for profit maximization.
Beyond the Promise of Democratization
The narrative around AI writing tools often centers on democratization: making eloquent expression accessible to everyone, leveling the playing field between experienced and novice writers, between native and non-native speakers. While these tools do offer certain benefits, helping non-native English speakers create professional communications and enabling small businesses to produce polished content, the story is more complex than simple democratization suggests.
Research reveals that the benefits of AI writing assistance are not equally distributed; studies show that AI provides greater efficiency gains for Western users compared to non-Western users, creating what researchers term "service quality harms," where certain groups must invest more effort to achieve similar benefits (Basu et al., 2023). This disparity raises fundamental questions about who truly benefits from these "democratizing" technologies and whose interests they ultimately serve.
While these tools make "good" writing more accessible, they are also making human writing skills increasingly optional. Why struggle with sentence structure when AI can fix it? Why develop your unique voice when AI can give you one that "works"? This reflects the broader logic of optimization capitalism: why invest in developing human potential when you can extract value more efficiently? The same mindset that strip-mines natural resources now strip-mines human creativity, packaging it into algorithms that can be sold back to us as a service.
Writing has always been more than just communication; it is thought made visible. When we struggle to find the right word, when we erase and rewrite a paragraph five times, when we articulate a complex idea, we are not just crafting sentences, we are developing cognitive muscles: critical thinking, problem-solving, creativity, and most importantly, the ability to translate abstract thoughts into concrete expressions. These muscles atrophy from lack of use.
Research in educational technology shows that excessive reliance on AI dialogue systems can negatively impact students' cognitive capacities, particularly when users accept AI-generated recommendations without questioning them (Liu et al., 2024). When AI handles the heavy lifting of writing, we lose opportunities to develop these fundamental skills.
Consider the difference between someone who has spent years developing their writing craft and someone who has primarily used AI assistance. The former has internalized the rhythms of language, developed an intuitive sense of flow and emphasis, and learned to trust their voice. The latter, despite producing technically competent text, may struggle with original expression when AI is unavailable. This dependency is not accidental; it is profitable. Just as planned obsolescence keeps us buying new phones, intellectual obsolescence keeps us subscribed to AI services. We are being trained to need what we once could do ourselves.
**
The Monoculture of Optimized Expression
**Perhaps most concerning is how AI is creating convergence toward a particular style of "optimal" writing. AI models are trained on vast datasets that embed certain biases and assumptions about what constitutes good writing. They tend to favor clarity over ambiguity, structure over experimentation, and efficiency over style. Recent research demonstrates this empirically: AI suggestions lead to diminished lexical diversity in writing, with users from different cultural backgrounds converging toward similar patterns of expression (Basu et al., 2023). However, whose definition of "optimal" are we accepting? The answer reveals something uncomfortable: AI writing tools reflect and amplify the values of corporate communication. They optimize for the kind of writing that moves products, closes deals, and maintains professional hierarchies. They are excellent at producing the language of capitalism: persuasive, efficient, and frictionless.