Verum

An AI tool that identifies bias and misinformation in political speeches and text. We aim to reduce confusion among voters and hope to create a more informed society for this upcoming 2024 election.

Comment

Mission

Empowering Informed Democracy with AI: Our mission is to provide voters with a steadfast tool that cuts through bias, inconsistency, and misinformation. By leveraging advanced AI technology, we guide users toward making confident and well-informed voting decisions for a more transparent and engaged democratic process.

What it does

Our platform offers a chatbot that aids individuals in examining political speeches and texts. Users have the option to either input a link to a transcript or article or select from our library of transcripts. Alternatively, they can also use their own text. After submitting the input, our tool forwards it to GPT-3.5 Turbo for a comprehensive assessment of bias. The analysis involves detecting bias and false information, offering explanatory interpretations, and presenting analytics for the percentage and count of biased statements detected in the text. Our platform is intended to assist users in recognizing partiality in political content and making knowledgeable decisions.

Story

Verum.AI's journey is a tale of serendipity and shared ambition, focused on crafting innovative solutions to promote inclusivity in design and large language models. We've united a talented group of designers and developers from both UC Davis and UC San Diego, forging connections both professionally and personally. Our front-end design was handled by Dereck Villagrana, while our back-end development team comprised of AJ, Marco, and Megan.

This harmonious blend of skills and experiences laid the foundation for Verum. Fueled by our shared vision, we embarked on a mission to create an AI solution that not only harnesses advanced AI technology to detect bias, inconsistency, and misinformation within political documentation but also places user needs and preferences at the forefront.

In conclusion, Verum embodies the successful synergy between designers and developers from UC Davis and UC San Diego, bound by a shared commitment to inclusive design and large language models. We're appreciative of the support and connections that have brought us to this point and remain dedicated to further enhancing our AI solutions for the benefit of all.

Challenges

One of the key challenges we encountered was enabling the analysis of both plain text and web content. To address this, we integrated an external API to extract text from external websites, making use of AWS Lambda to orchestrate these processes. The core of our system was the integration of GPT-3.5 Turbo, an AI model, to perform bias and accuracy analysis. Managing the prompt input for GPT-3.5 Turbo was a task in itself, as we needed to dynamically summarize text to fit within prompt limits while preserving the context. Throughout this journey, we tackled various technical hurdles, including optimizing API calls and handling asynchronous processes. However, the end result was a powerful and versatile system that could analyze text and web content for biases and inaccuracies, ultimately providing users with valuable insights into the information they encountered online. Our journey from concept to final product was marked by innovation, adaptability, and a commitment to delivering a robust solution in the realm of AI-driven content analysis.

What's next for Verum

Our team is dedicated to our mission of creating the platform for the upcoming 2024 election. As we enter the new year, we urge more individuals to utilize the platform as more politicians voice their perspectives. Our end objective is to boost voter trust and cultivate a more knowledgeable society.

Built With

Share this project:

Updates