Inspiration
There is lots of misinformation around the world created by humans and AI. We wanted to create a tool that allows users to seamlessly see if some text they see is truthful or not.
What it does
This AI parses through the text and analyzes each claim by searching up trustworthy sources and comparing information in those sources compared to the given source.
How we built it
We used bedrock, textract, and comprehend from AWS in order to parse the claims made in a text and be able to compare those claims to claims in other sources on the internet from more trustworthy sources.
Challenges we ran into
There were extreme challenges with choosing what backend to implement, as some of AWS things like Kendra weren't working. We had to pivot a lot to make the backend work.
Accomplishments that we're proud of
We made a solid frontend and backend that are interconnected. We have a product that we can show, even though it requires lots of improvements.
What we learned
We learned how to use AWS and its features. We also learned how to use Cursor and how to use it to speed up our workflow
What's next for TruthLens
We hope to improve the backend model to more accurately identify what is misinformation and what isn't in the future.
Log in or sign up for Devpost to join the conversation.