SafePath AI

SafePath is an AI-powered chatbot that helps undocumented high school and college students get basic legal info and emotional support.

Comment

Inspiration

As a Hispanic first-generation student directly impacted by current immigration laws, I’ve developed a strong passion for helping others—especially students who may feel scared, isolated, or uncertain simply because they don’t know where to turn. Growing up in a large family, I often found myself in that position, and I know how important it is to feel seen and supported. SafePath AI is a deeply personal project. While it may not change the larger systems at play, my hope is that it can provide comfort, clarity, and access to basic information for students who are afraid to ask or unsure of where to begin.

How we built it

SafePath AI is an AI-powered chatbot designed to support undocumented students and others in similar situations. It offers answers to basic legal questions related to immigration and provides empathetic, emotionally supportive responses to help users feel heard and less alone. It serves as a judgment-free space where users can ask sensitive questions without fear.

SafePath AI was built using:

Frontend: React for the user interface, styled with Tailwind CSS

Backend: Flask to handle API requests and communicate with the AI model

AI Integration: GROQ API to generate intelligent, context-aware responses

Hosting: Local development environment during the hackathon, with plans to deploy via Render or Vercel

Version Control: Git and GitHub for collaboration and project tracking

Challenges we ran into

  • Prompt engineering: Crafting prompts that balanced factual legal information with emotional sensitivity was tricky. I had to iterate multiple times to avoid generic or unhelpful responses.

  • Data sensitivity: Finding a way to serve undocumented users without collecting personal data was important, so we kept everything anonymous and focused on general questions.

  • Balancing tone: Making the chatbot feel supportive and human without making it seem like a replacement for therapy or legal advice was a constant balancing act.

Accomplishments that we're proud of

  • Designing an empathetic, accessible, and user-friendly interface

  • Building a project that reflects a real need in our community

  • Staying focused on user experience for a sensitive topic

  • Learning how to integrate multiple tools and frameworks efficiently

What we learned

  • How to use AI responsibly when dealing with sensitive subjects

  • The importance of user-centered design in emotionally charged use cases

  • How to manage time and scope effectively during a hackathon

  • That personal projects can have powerful impact when tied to real-world problems

What's next for SafePath AI

  • Mobile Version: Create a mobile-friendly app for easier access

  • Resource Integration: Link users to local support organizations, helplines, and legal clinics

  • Deployment: Host it on a public platform so it can reach real users who need it

Built With

Share this project:

Updates