An app that captures how often women get to speak at events

Women’s voices are frequently marginalised in organisations and at conferences due to innate biases. Recognising this, IT solutions company Thoughtworks has developed an AI-powered tool to quantify these biases.

• Also read: How India’s first woman mining engineer cut a rocky path for herself

The Gender Gap App (G-App) measures disparities across four key dimensions: political, educational, economic and physical integrity. By analysing data from various sources, the app reveals hidden inequalities that traditional metrics might miss. This AI-based innovative tool has already been piloted within a UN organisation and is now ready for wider deployment.

The app was developed in collaboration with Women at the Table, a Geneva-based organisation focused on advancing gender equality. This partnership brought together expertise in technology and advocacy, resulting in a tool that identifies biases and empowers organisations to take targeted action.

Collaborative effort

It took two women to identify the importance of measuring the biases and finding a solution to measure them. When Caitlin Kraft Buchman, Founder and CEO of Women at the Table approached Rebecca Parsons, the Chief Technology Officer of Thoughtworks, she loved the idea and connected Buchman to her team in India to develop the app.

“We wanted to dig deeper from just sheer numbers to representation and participation, active participation,” explains Buchman. In other words just counting the number of women in a room doesn’t measure equality. The idea is to also measure and report on who gets to speak and who gets heard at world-shaping events.

The app does just that diving into the dynamics of these gatherings. The app analyses both structured data about the conference (attendees, gender, roles) as well as unstructured data from audio and video recordings. This allows it to measure not just speaking time, but also the content of speeches, using machine learning algorithms to identify topics and assess influence.

Nuanced understanding

 “Even though these international assemblies often [have] 60 per cent men to 40 per cent women...almost across the board, men were speaking 80 per cent of the time and women speaking just 20 per cent,” Buchman told businessline over the phone from Geneva.

By quantifying the hidden biases, the app promises to catalyse meaningful change in organisations worldwide. It provides a more nuanced understanding of women’s participation and influence and has the potential to drive significant change in international forums.

The tool comprises the processing of two sets of data, a visualisation engine and a topic modeling algorithm. The anonymised delegate information and transcripts - from formal speeches as well as chamber discussions - are processed and matched to topics at the conference, which the algorithm generates.

“We were able to bring some machine learning tools and algorithms to analyse what was spoken, the transcripts of what was spoken, and then tie that back to the themes of the conference itself,” Satish Vishwanathan, Head of Social Impact at ThoughtWorks India, explains.

This reveals whether women are being relegated to ‘soft’ topics like gender and health, or whether they are participating in discussions on ‘hardcore’ issues like finance and climate change.

The app has been piloted at several high-profile events, including the Paris Peace Forum and the UN Conference on Trade and Development and discussions are underway for its wider adoption within the UN system. While the tool’s primary focus is on gender, it can also be used to analyse other forms of disparity, such as age or nationality.