Inspiration
The inspiration of SafeSpeak was to combat online sex trafficking and protect vulnerable individuals, especially those using messaging apps. We wanted to create a way to analyze and combat sex trafficking by creating an analysis messaging application that allows users to determine how likely sex trafficking may be involved.
What it does
SafeSpeak is a safety-focused messaging platform enhanced with a sex trafficking detection algorithm. When users chat, the system will automatically analyze messages and flags conversations that show potential indicators of trafficking behavior. If it detects any signs, the app will warn users on the situation.
How we built it
We utilized React for the frontend interface, Python (Flask) as the backend API, a machine learning model to detect trafficking-related language patterns, as well as Heroku to host the service.
How it works
In this example we can see two users, Alice and Bob. Bob wants to send a message to Alice. Great! His message is encrypted and sent straight to Alice. Alice receives the message and it is only decrypted on her device. After viewing the message, she has an option to run an analysis on her conversation with Bob. Safespeak will securely run the analysis on her device. There will be no server contact, this analysis is only for Alice. That is, unless the conversation with Bob is extremely abusive. Then it will be sent to SafeSpeak servers where an expert can reach out to Alice.
Challenges we ran into
- Linking the front and backends.
- Training the AI model.
- Setting up Heroku to run with our code while also using a public domain.
Accomplishments that we're proud of
- Successfully integrating live detection into the messaging app.
- Building an app serves for a meaningful cause.
- Creating a user-friendly interface.
What we learned
- How to utilize different API calls between front and backend to obtain specific pieces of data.
- ML, web development, and API integration.
- The importance of balancing privacy, safety, and performance in real-time communication apps.
What's next for SafeSpeak
- Improving the accuracy and robustness of the detection model.
- Adding multi-language support.
- Partnering with anti-trafficking organizations for resource integration.
Log in or sign up for Devpost to join the conversation.