Inspiration
Develop this app to empower people to support one another. Sometimes, those in need may feel afraid to ask for help, but with AI, we can detect and predict when someone might be in trouble. This way, others can step in and offer assistance when it's most needed.
What it does
The app enables users to share thoughts and updates, or fetch data from user's social media, while Gemini AI analyzes posts to detect any signs of potential violence. Each update is assigned a risk score, and if a high-risk situation is identified, nearby users are alerted based on location to offer timely support.
How we built it
I developed the app using React for the frontend and Node.js with Express.js for the backend. The Gemini AI model (gemini-1.5-flash) powers the content analysis, enabling real-time risk assessment. The entire app is deployed on Render for reliable access and scalability.
Challenges we ran into
Our initial plan was to pull new feeds from social media platforms such as Instagram and Facebook to analyze and detect potential risks using the Gemini AI model. However, due to privacy restrictions on the Facebook Developer Platform, we were unable to fetch new social media feeds directly. This limitation forced us to rethink our strategy and focus on in-app updates shared by users, ensuring that we still have the necessary context to detect distress signals while adhering to privacy standards.
Accomplishments that we're proud of
We successfully integrate generative AI to analyze users' updates. This allows us to calculate a risk score based on the content of their posts and display it in the app's dashboard in real-time. By leveraging the power of the Gemini AI model, we’ve been able to provide users with valuable insights into their potential risk level and alert them to any danger they might face.
What we learned
We learned how to build and deploy an app from scratch, explored how generative AI APIs can be applied, and understood that AI requires continuous fine-tuning to achieve accuracy.
What's next for Safety predictor
The next features we’re working on is enabling users to send direct messages or emails to other users when they observe something concerning or when a potential risk is detected. This feature will allow users to take immediate action by reaching out to others and alerting them to potential danger in real-time, creating a more proactive safety network.
We believe this functionality will enhance community collaboration and further empower users to help one another when it's most needed, making the app even more effective in preventing violence and ensuring a safer environment for everyone.
Log in or sign up for Devpost to join the conversation.