How Automated Writing Tools Are Making Everyone Sound American

ITHACA, N.Y. — Every time you accept that gray text suggestion in Gmail or let predictive text finish your sentence, you might be unwittingly adopting a more generic, Western writing style. That’s especially true if you’re from a non-Western culture, according to eye-opening research from Cornell University.

The new study shows how AI writing suggestions push Indian users toward American writing patterns and provide fewer productivity benefits to Indian users compared to their American counterparts. The implications stretch far beyond mere word choice – these technologies may be slowly homogenizing global cultural expression.

The Cultural Cost of AI Writing Assistance

When Indians use AI writing assistants, they not only adopt more Westernized vocabulary choices but also change fundamental aspects of their writing style, like using different sentence structures and descriptions that reflect Western viewpoints rather than Indian cultural perspectives. For example, when describing traditional Indian festivals with AI assistance, users produced more generic descriptions focused on “family gatherings” and “celebrations” rather than including specific cultural elements like religious rituals or regional variations.

The research team – Dhruv Agarwal, Mor Naaman, and Aditya Vashistha – conducted experiments with 118 participants split evenly between India and the United States. Participants completed culturally-grounded writing tasks with topics ranging from describing favorite foods and celebrities to festivals and formal workplace emails. Half received autocomplete suggestions from GPT-4o as they wrote, while the others wrote without AI assistance.

Hidden Biases in AI Suggestions

By comparing writing from all four groups, researchers identified clear patterns of cultural homogenization. With AI assistance, Indian writing became significantly more similar to American writing patterns.

The initial AI suggestions were consistently Western-biased. The first food item suggested was always pizza or sushi, and the first suggested festival was invariably Christmas, regardless of what the user was trying to write about. These subtle nudges steered writers toward Western cultural references.

What’s particularly concerning is that many people don’t realize this influence is happening. Most modern digital writing tools (Gmail, Google Docs, Microsoft Word) have AI suggestions enabled by default, silently shaping how billions of people express themselves.

Beyond the cultural implications, the researchers discovered that AI writing tools provided unequal benefits. While Americans saw a roughly 30% increase in writing speed with AI assistance, Indians experienced smaller productivity gains despite accepting more AI suggestions. This raises questions about whether these technologies truly serve global users equitably.

Measuring Cultural Homogenization

To measure the extent of this homogenization, researchers trained a machine learning model to distinguish between Indian and American writing. When analyzing texts written without AI, the model achieved 90.6% accuracy. However, when examining AI-influenced writing, accuracy dropped to 83.5% – clear evidence that cultural differences were being erased.

This wasn’t just about vocabulary or grammar corrections. The AI influence extended to deeper aspects of writing, including lexical diversity (the variety of unique words used). Indians naturally used fewer unique words in their writing compared to Americans, but with AI assistance, their patterns shifted to match Western styles.

For example, when Indian participants wrote about traditional biryani without AI, they included details about regional variations and specific accompaniments. With AI, these cultural specifics disappeared, replaced by generic descriptors like “rich,” “flavorful,” and “aromatic” – subtly reframing Indian food through Western perspectives.

The researchers describe this as a form of “AI colonialism,” where Western-centric AI models inadvertently reinforce Western cultural dominance by erasing distinctive cultural expressions. As these technologies continue expanding globally, this subtle erasure could have profound implications for cultural diversity worldwide.

“This is one of the first studies, if not the first, to show that the use of AI in writing could lead to cultural stereotyping and language homogenization,” said senior author Aditya Vashistha, an assistant professor of information science, in a statement. “People start writing similarly to others, and that’s not what we want. One of the beautiful things about the world is the diversity that we have.”

Writing reflects thinking and identity. When AI suggestions push everyone toward Western patterns, they risk creating what the researchers call “cognitive imperialism” – the dominance of a single cultural perspective.

“These technologies obviously bring a lot of value into people’s lives,” said first author Dhruv Agarwal, a doctoral student in the field of information science. “But for that value to be equitable and for these products to do well in these markets, tech companies need to focus on cultural aspects, rather than just language aspects.”