AIcebreaker

a web app that anonymously connects two users through chat based on interests and prompts them with icebreakers generated with a natural language processor

Comment

Inspiration

For this hackathon we wanted to create a web app. Since there has been so much hype around AI with the release of ChatGPT, we believed creating a chat app would be the perfect way to make AI useful to the mainstream. There has also been a lot of concerns around the use of AI and we wanted to change the rhetoric around this discussion by showing that AI can be used to bring people together.

What it does

AIcebreaker is a web app that anonymously connects two users through chat based on interests. It spurs conversation by posting AI-generated prompts, or icebreakers, related to users’ shared interests. Users are first directed to the landing page where they specify interests to help with matchmaking through a tagging system. Then, the website pairs together two users and directs them to a chat window where a custom icebreaker is created using natural language processing. Within the chat, users can both opt to generate another question, or disconnect and chat with someone else.

How we built it

We used the reactjs framework and javascript to build the frontend and backend, and for the natural language processing portion of generating text, we used GPT3 API.

Challenges we ran into

With only 24 hours, there was definitely a restraint to how much we could accomplish as there was so much we wanted to implement that we were not able to. We were also not so familiar with the framework we were using and had to mess around with it before we knew how it worked. With so many different ways of implementing our ideas, we had to prioritize certain functions over others and were not able to implement every function we wanted to.

Accomplishments that we're proud of

We were able to generate prompts that were interesting and relevant to the interests of different people, and we were able to make a demo that shows a prototype of what our application could be.

What we learned

We learned to use new frameworks and API, and navigate a project.

What's next for AIcebreaker: Streamlined UI/UX design, scalable matching algorithm, the option to create an account and save your conversations, waiting room for people without a match, random matchmaking

Share this project:

Updates