The research

Gendered Speech, Gendered Machines

Gender aspects of human–AI communication

Women adopt generative AI less than men, and the gap is often explained by access, networks and confidence. I’m looking at something else: how the way we’ve learned to speak (hedging, politeness, directness, who asks and who tells) shapes what happens when we talk to a machine. Not “how women prompt”, but how the conversation itself works for them.

Independent researcher since 2025 · aiming to start a PhD in 2027.

Follow the research

Occasional updates on the project, the PhD, and my search for a supervisor — a few emails a year, no more.

The questions

RQ1

Are there measurable differences in how women and men structure their communicative acts with AI — hedging, directness, how requests are framed, and how they position themselves in the conversation?

RQ2

How far are those differences predicted by gender-structured social norms, as opposed to technical literacy or other demographics?

RQ3

What do the answers mean for AI design, inclusivity and adoption — so tools do not quietly reward one way of talking?

The Unhyped Shelf

AI papers worth your time, minus the hype. One link to the open paper, a few paragraphs of what I think, and what I’m still unsure about.

Notes from the research

Nothing here yet. Notes are coming soon.

Get the Unhyped Shelf by email

When I publish a new note on a paper worth your time, I’ll send it to you. Nothing else.

Training & certifications

Anthropic logoAnthropic
UNDP Human Rights Impact of AI AssessmentHuman Rights Impact of AI Assessment (HRIA) Bootcamp · UNDP
Moscow State University emblemMSU Linguistics and Teaching

Research FAQ

›What is Maria Nightingale researching?
Gender aspects of human–AI communication: whether women and men structure their conversations with generative AI differently (in framing requests, hedging, directness and politeness), and what that means for how useful AI feels to them. The working title is Gendered Speech, Gendered Machines.
›Is there a gender gap in AI use?
Studies report a persistent gap of around 20% in generative-AI adoption between women and men, usually explained by access, networks and workplace factors. This research asks whether communication styles play a part too. (Source: a meta-analysis across 76 sources and 100+ countries, Harvard Business School Working Paper 25-023, 2024.)
›Why does it matter?
If AI tools quietly reward one way of talking, they can widen existing gaps at work and in education. Understanding that helps design fairer tools and better AI training.
›Where will the PhD be?
In France, starting 2027, in Information and Communication Sciences. (No institution named.)
›Can I collaborate or take part?
Yes, write to [email protected].