DiversitAI

• 

  printed results

• 

  video analysis dashboard

Inspiration

To improve inclusive language at scale by detecting non-inclusive language and making suggestions. The goal of this project is to create a culture of belonging by promote inclusive language establish a healthier inlcusive culture. I identified this issue and made the decision to develop a solution after observing honest mistakes being made by other coworkers or myself in our prior jobs while utilizing vocabulary that was not particularly inclusive. Most mistakes happen out of lack of information thus we decided on the approach we are using of educating not shaming.

What it does

I realized that mistakes happen in real life. With most meetings happening on platforms like zoom and Teams, it would be helpful to get an analysis of your meeting after it’s done. Thus, I came up with the idea of analyzing videos with firstly flagging non-inclusive words. then detecting things about the meeting like profanity count, non inclusive words count and making suggestions. for instance if the user says the word "guys", it suggests "folks"

How we built it

thanks to AssemblyAI API’s, I am able to get the transcript of any video/audio and convert it to speech. Further I was able to extracts the dominant speaker, the amount of non-inlcusive words, speakers count, profanity count and generate an alternative for the words.

What's next for DiversitAI

User Testing To implement the project and actually publish it. Crowdsourcingg, having users input suggestions of triggering words. Make more people aware about using inclusive language. Implement on more platforms

Built With

• css3
• html5
• javascript
• python