Power Dynamics in Meetings

Gender is another powerful driver of meeting dynamics, shaping who gets heard, how contributions are interpreted, and whose ideas ultimately influence decisions.

Gender shapes how people show up in meetings more than we often realize. Across decades of research—from faculty meetings to scientific conferences—men speak earlier, speak more, and dominate Q&A sessions even when panels are gender-balanced. That creates a familiar challenge for women, because speaking time often gets interpreted as confidence, status, or leadership—advantages that men are more likely to be granted automatically. Researchers call this the “Babble Hypothesis”: we routinely mistake talking more for leading. In one study, every extra 39 seconds of airtime earned someone another “leader” vote, no matter what they actually said.

But when teams use AI, our data suggests the dynamic starts to flatten. In our dataset, women contribute about 9% more airtime than men relative to their representation in the meeting. One likely reason: when people know their words are being captured, summarized, and potentially revisited, that ambient awareness can make participants more reflective about how much they are talking in the meeting. It’s a modern-day Hawthorne effect and one that disproportionately benefits women, whose contributions are more likely to be interrupted, discounted, or overwritten in traditional meeting dynamics. As well, when talk-time patterns are measured and visible, leaders and facilitators can see in real time who’s authentically contributing to move the conversation forward and who’s not, and adjust.

Historically, studies point to faster talk speeds as a frequent male trait, possibly because of confidence and a desire to convey more information. But when teams are freed from the distraction of manual documentation, everyone can focus entirely on the content, leading to more dynamic, free-flowing, and faster conversations, with more contributions from participants overall. 

Where we continue to see clearer divergence is in language. In many organizations, men use dismissive or non-inclusive language (such as “mansplaining”) more frequently, while women receive less credit for similar ideas. In our dataset, we see this pattern clearly: women use significantly fewer non-inclusive terms (1.7 per person per meeting on average, compared with 2.2 for men). This is consistent with research showing that women tend to rely more on facilitative, connective language that prompts participation and keeps conversations moving. Unlike speaking time or pace, inclusive language is difficult to self-correct in the moment; it reflects years of habits and social conditioning. That may be why we don’t see the same level of flattening here as we do in other dynamics.

Some of the clearest signs of inequity aren’t about who speaks—they’re about who stops speaking. Read AI captures “ghost mode”: moments when someone stays off-camera and muted, a reliable signal that they’ve pulled back from the conversation. In our data, women enter ghost mode 19% more often than men. That gap likely reflects the added cognitive and social burden of constant self-monitoring—sometimes described as the mirror effect—and the extra effort required to remain engaged.

Hybrid work has introduced a new layer of power dynamics into meetings. People in a physical conference room enjoy various "proximity biases." They often speak earlier, more often, and with more ease. They benefit from micro-interactions remote colleagues never see: the pre-meeting hallway chat, the side glance that cues a handoff, the shared laughter that warms up the room, and the nonverbal signals that help them time their entry into the conversation. 
‍
In our dataset, we see these imbalances clearly. Conference-room participants speak more than five times as much as remote participants (after normalizing talk time by the number of participants in each location). This is the largest gap across any demographic dimension we analyzed.
‍
Why do hybrid meetings show the most extreme power imbalances? When people know they’re being observed—even subtly by AI analytics in virtual meetings—they are likely to regulate their behavior: sharing airtime more evenly, pausing before jumping in, and staying aware of who hasn’t spoken. We can see this effect most clearly in hybrid meetings, where some participants remain visible through analytics while others sit in physical rooms without the same cues. 

We also see imbalances in speaking pace and participation style. In meetings captured by Read AI, people in the room speak faster (about 181 words per minute compared with 172 for remote colleagues), which can make it more difficult for remote participants to jump in. They also ask nearly twice as many questions per meeting (6.2 versus 3.7 per meeting, on average) and use more filler words (38 words versus 24 words per meeting), both signs of conversational comfort and dominance. Together, these behaviors make it even more difficult for remote attendees to break into the discussion or steer it once it’s underway.
‍
Language patterns tell a similar story. In-room participants use more non-inclusive terms than remote colleagues (2.7 versus 1.9 non-inclusive words per person per meeting). A likely reason is that people in the room feel more comfortable and less scrutinized. They can read the room, gauge reactions, and recover more easily if something lands poorly.

AI gives teams a way to counteract that drift. By surfacing real-time gaps in talk time, ghost-mode behavior, speaking pace, and acknowledgment patterns, leaders can intervene before remote participants fade out of the discussion.

Cognitive and communication differences shape meetings in subtle but important ways. Traditional meeting formats privilege a certain set of behaviors: fast verbal processing, rapid-fire turn-taking, constant camera presence, and the ability to think out loud. That setup works well for people who thrive in spontaneous, high-tempo environments—but it works far less well for employees who process information differently, including those with ADHD, autism, dyslexia, sensory sensitivities, introverts, reflective thinkers, non-native speakers, or visual processors.

What leaders often interpret as “quiet,” “hesitant,” or “disengaged” is often something else entirely: people operating at a different cognitive cadence. A pause isn’t uncertainty. A slower speaking rate isn’t hesitation. A preference for chat over verbal input isn’t detachment. Without visibility into these patterns, the format of the meeting—not the quality of ideas—determines who gets heard.

The cost of this is real. Many of the strengths associated with neurodivergent or reflective thinkers—pattern recognition, scenario analysis, first-principles reasoning, creative problem-solving—can significantly improve team performance, but only if meeting structures make space for those contributions to surface.

AI-powered meetings can help leaders surface these hidden patterns. Insights gleaned from meeting interactions can reveal signals such as:

• A contributor who enters the discussion only after several others have spoken.

• A colleague who switches from verbal comments to chat as speaking rates accelerate.

• A participant whose engagement drops when agendas or materials aren’t shared in advance of the meeting.

Decades of organizational research have made one thing clear: meetings are shaped by power dynamics. Hierarchy, gender, proximity, and cognitive style all influence who speaks, who gets heard, and whose ideas shape outcomes. What is new is leaders’ ability to see these dynamics clearly, consistently, in real-time, and at scale.

The table below brings together what research has long shown about power dynamics in meetings with what actually happens in meetings analyzed by Read AI. When participation is made visible with AI, long-standing power dynamics tied to formal status and gender begin to flatten. But in hybrid meetings, where it’s easier to overlook the presence of AI, proximity-based imbalances persist.

These insights underscore a broader point: meetings aren't just about showing up. They are environments where power, participation, and performance interact in measurable ways. Teams that understand how real contributions unfold—and create norms that surface them—are the ones performing at the highest level today.

With AI as a partner, leaders can take meetings from invisible, habit-driven interactions to deliberate systems that surface the best ideas, strengthen decisions, and drive real business results.