NewsCheckerAI

What is NewsCheckerAI

NewsCheckerAI is an AI-powered tool designed to help users identify misinformation and verify the accuracy of news headlines they encounter online. By leveraging advanced artificial intelligence models, NewsCheckerAI analyzes headlines and provides instant predictions with a confidence score, allowing users to make informed decisions about the credibility of the information they consume. Supporting both English and Turkish headlines, this user-friendly application aims to combat the spread of misinformation and promote a more informed and trustworthy digital environment.

Download the Extension

How It Works?

NewsCheckerAI is an open-source, AI-powered application designed to verify the credibility of news headlines. At its core, the application uses a Logistic Regression model trained on a dataset of 20,000 fake or satirical news headlines and 20,000 real news headlines. This dataset was meticulously prepared to ensure a balanced representation of both categories, enabling the machine learning model to learn the nuances that distinguish real news from misinformation.

The training process involved using advanced natural language processing techniques, where each headline was converted into numerical features using methods like tokenization and vectorization. These features were then used to train the Logistic Regression model, optimizing it to predict whether a given headline is real or fake with high accuracy.

Once trained, the model is deployed to the backend, where it can be accessed via an API. When users input a news headline, the backend processes the text and sends it to the machine learning model for prediction. The model returns a confidence score and prediction, which is then displayed to the user in real time.

NewsCheckerAI also supports both English and Turkish headlines, making it versatile and accessible to a wide range of users. The open-source nature of the app ensures transparency and encourages contributions from the developer community to enhance its functionality and accuracy.

FAKE

REAL

To Do

Model Selection: Users will have the ability to choose from multiple AI models based on performance, accuracy, or specific needs. This will allow them to tailor their experience to their preferences or system capabilities.
Custom Model Upload: Users will be able to upload their own trained models into the application. These models can be used personally or shared with others, fostering a collaborative community of AI enthusiasts.
Model Marketplace: A sharing option will be introduced, where users can make their models available to others. Model creators can even monetize their contributions, allowing them to generate income based on the popularity and usage of their models.
Web3 Integration: We are planning to develop a specialized AI model tailored to the needs of the Web3 world. This model will focus on analyzing headlines and news related to blockchain, cryptocurrencies, and decentralized applications to combat misinformation in these emerging fields.
Community Contributions: As an open-source project, we aim to encourage community involvement. Developers will be able to contribute to the codebase, enhance features, or even propose new AI models for integration.

2023 June

Machine Learning Models

In June 2023, we developed our initial machine learning models, including Linear Regression and three additional models, laying the foundation for our project.

2024

Research and Testing

Throughout 2024, we rigorously tested our models across multiple languages, benchmarked them against industry-leading models, and showcased our progress at the IEEE conference.

2025 January

Product Launch

After extensive testing and refinement, we launched our API in January 2025, delivering the model as a Minimum Viable Product (MVP) to meet real-world demands.

TBA

Advanced Model Development

We plan to enhance our models further by incorporating Web3 content, driving innovation and adaptability for decentralized applications.