FactFlash

We've had enough of politicians lying boldly on live TV.We've had enough of emotional manipulation instead of logical arguments. Say hello to factflash: uncovering fake news and fallacies in real time

Comment

Inspiration

In an era where misinformation spreads rapidly, especially in political speeches and public debates, the need for real-time fact-checking and logical analysis has never been greater. The inspiration for factFlash stemmed from the desire to create a tool that empowers individuals to critically evaluate the information they consume, ensuring that they can distinguish between truth and falsehood, and logical arguments and fallacies.

What it does

factFlash continuously monitors the audio from another Chrome tab and analyzes the claims and logical arguments made within it. It identifies and flags fake news, misinformation, and logical fallacies in real time, providing users with instant feedback on the content they are listening to. This is particularly useful for vetting political speeches, educational debates, and any content where the integrity and logic of the information are paramount.

How we built it

FactFlash uses React for frontend application interface, an express server on the backend, and OpenAPIs GPT and Whisper models for transcription and analysis. The frontend continuously sends the audio to the backend, which transcribes, analyzes, and sends the analysis back to be displayed on the React app.

Challenges we ran into

The major challenges we faced were solving how to reduce API call frequency, perfecting the GPT prompt to accurately detect fallacies and fake facts, and integrate everything into a working prototype.

Accomplishments that we're proud of

We are proud to have developed a functional prototype of factFlash that can accurately monitor and analyze live audio streams for misinformation and logical fallacies that can work with any browser tab with audio playback. It's already a big step in trying to solve the problem we aspired to fix.

What we learned

Throughout the development of factFlash, we learned the importance of balancing accuracy with real-time performance in audio and text analysis. Moreover, we learned about the various types of logical fallacies and misinformation tactics commonly used in speeches and debates.

What's next for factFlash

Moving forward, we aim to enhance factFlash by reducing our latency, adding character identification in the transcription service, and improving our audio clipping logic to make analysis more natural.

Built With

Share this project:

Updates