BALTIMORE — Johns Hopkins study ties AI tools to women’s writing tone

A study conducted at Johns Hopkins University, as reported by Moneywise.com, found that artificial-intelligence editing tools may make women's workplace writing read as less professional than men's after the software rewrites or suggests edits to the same material. The finding adds to a growing body of workplace research examining how AI writing assistants interact with existing gender dynamics in professional communication.

What did the Johns Hopkins study find?

According to the Moneywise.com report, the study examined how AI tools handle workplace writing produced by men and women and concluded that the software's suggested edits tended to shift women's text toward phrasing that readers judged less professional than men's edited text. The underlying research paper, methodology details, and sample size were not specified in the available source material, so this report does not state them. Readers seeking the full study design should consult the originating Moneywise.com article and the underlying university research when it is published in full.

How do AI writing tools actually edit text?

Workplace AI writing assistants — the category that includes grammar checkers, tone-adjustment features in email clients, and standalone drafting tools — operate by comparing a user's draft against patterns learned from large volumes of training text, then suggesting word choices, sentence structure, or tone shifts intended to match those patterns. Because those tools apply the same underlying model to any input, a difference in how the output reads by the writer's gender points to a difference in the input text, the model's training data, or both — a mechanism the Moneywise.com report does not fully isolate based on the material available.

Why might the effect differ by gender?

The source material does not supply a tested explanation for why AI editing might affect men's and women's writing differently, and this report does not speculate beyond what Moneywise.com and the Johns Hopkins study reported. Broader research on AI in professional settings has flagged related dynamics. A separate analysis highlighted by HTT News on the emotional dynamics of AI in the workplace found that how employees relate emotionally to AI tools — trust, reliance, and perceived judgment — shapes how those tools get used day to day, a factor that could compound with any tone-related effects in editing software.

What should employers and employees using AI tools do?

For companies that have adopted AI writing assistants in email, performance reviews, or external communications, the practical question raised by the Johns Hopkins finding, as described by Moneywise.com, is whether tone-suggestion features should be applied uniformly or reviewed for consistency across writers before being sent. The study's reported conclusion does not come with a specific corrective recommendation in the available source material, so none is asserted here beyond the general principle that any automated edit affecting professional tone warrants a human read-through before it goes out under an employee's name.

Some consumer-facing AI tools market themselves on helping users keep a consistent voice across drafts rather than having the software substitute its own tone. Smorgi Apps' AlterEgo and mindfulAI Keyboard are examples of tools built around preserving a user's own phrasing while assisting with edits, an approach that sidesteps the blanket-rewrite model the Johns Hopkins study appears to be examining. Smorgi Apps has also said its forthcoming Artiglio tool is in development, though it has not yet shipped.

What is still unknown about the study?

The Moneywise.com report, as supplied in the source material for this story, does not include the Johns Hopkins study's full citation, publication venue, peer-review status, or the researchers' names and institutional titles. It also does not specify which AI tools were tested, how professionalism was scored, or how large the measured gap was between men's and women's edited writing. Those gaps mean this report describes the headline finding attributed to Moneywise.com and Johns Hopkins without independently verifying the study's design. Readers who want the underlying numbers should look for the original Johns Hopkins publication once it is available in full.

The broader pattern — AI tools reshaping how professional writing reads, and doing so unevenly across users — fits into ongoing workplace research on AI adoption, including the emotional-dynamics findings referenced above. Employers evaluating AI writing software for company-wide use may want to test outputs across a range of writers before standardizing on any single tool.

Artiglio is A coming-soon iPhone chief of staff for briefings and drafts. Not on the App Store yet.