Factly

Inspiration

The inspiration comes from the concern of "Fake News" and the spread of misinformation in the current climate of not only the

What it does

It takes an article, gives the reader a percentage of how reliable the news is, and allows the reader to decide what to do with that using an OpenAI API key, which the article parses statements to make a conclusion

How we built it

We created a Chrome extension that allows the User to get a response at the click of a button.

Challenges we ran into

We ran into challenges with LLMs, such as Ollamma not having internet access and not giving a recent take on the biased article, and Gemini, for which we originally developed a chutes and ladder model that assigns a truth value to the LLM and cross-references them to reduce bias, but we ended up scrapping that due to the LLM being restricted in terms of political stance. and lastly we ran into the challenge of hosting the backend on a server, and we fixed this by changing from python to javascript primarily in our code base

Accomplishments that we're proud of

• Creating a Chrome extension that has a simple UI does not allow for much confusion
• Learning how to work as a team, delegate tasks, and collaborate with others

What we learned

We learned how to collaborate as a team Problem-solve and iterate plan and provide an excellent overall framework for the idea Put into perspective the User's needs and wants when it comes to a product like this

What's next for Factly

The implementation of OpenCV and more work on the Chutes and Ladders model for the future to reduce LLM-specific bias and have an algorithm that can be open-sourced and potentially be iterated into databases for other companies

Built With

• chrome
• css
• html
• javascript
• openai
• python