Detecting Fake News Using AI in Liberia

Local Chapter Liberia Chapter

Coordinated byLiberia ,

Daikukai Bindah

Status: Completed

Project Duration: 01 Jul 2023 - 26 Aug 2023

Open Source resources available from this project

Github Repository

Project background.

In this digital age, the spread of fake news has become a significant challenge. Many years ago, such information was carried to its target audiences through word of mouth, but today the spread of misinformation has been hastened and weaponized by social media, which is one of the most convenient ways of conveying information in Liberia. From gruesome experiences, fake news carried throughout Liberia has led to atrocities and nearly brought the nation to the brink of collapse. The action of some Liberians has set in motion negative communal responses that have unleashed chaos in the past. Fake news in Liberia is commonly called “Dey Say,” a statement told by an unknown person. Dey Say is usually accepted as fact among many young Liberians and illiterates.

In 2020, violence took place in Gbarpolu County during the senatorial elections. The election was fueled with hate speech and misinformation, which spread throughout the nation via social media. The 2020 senatorial elections in Gbarpolu hang in uncertainty after a local paramount chief in Normondatonu district, called Chief McGill Wleh, confiscated the ballot boxes on the morning of the elections, leaving voters stranded. Chief Mcgill Wleh’s justification was that he believed citizens from Sierra Leone planned to cross the border and vote in the senatorial elections – information that has not been proven to be factual. The National Elections Commission (NEC) tried to organize a rerun, but that plan was aborted following more violence during which supporters of independent candidate Kanneh Botoe and locals in Normondatonu clashed in multiple acts of violence that targeted NEC staff as well.

The problem.

Fake news poses a severe threat to the integrity of information dissemination in Liberia. With the increasing availability of social media platforms and the rapid sharing of information, distinguishing between authentic news and fake news has become increasingly challenging for the general public.

Liberia, like many other countries in West Africa, has faced detrimental effects of misinformation and disinformation, which have led to social unrest, political instability, and public mistrust. Given that the nation is preparing for its national elections in October 2023, there will likely be a surge of misinformation from politicians in the media. This has the potential to disrupt Liberia’s democracy and plunge the nation into a crisis similar to the one it suffered 33 years ago. To combat this problem, there is a need to leverage the power of artificial intelligence (AI), which can be a transformative solution by using AI algorithms and machine learning techniques.

Consequently, misinformation can spread rapidly, causing confusion, divisiveness, and even harm to individuals and communities as seen in the case of the 2020 senatorial election. The need for an AI-powered solution arises to alleviate the burden on individuals to manually fact-check and verify news articles by harnessing the capabilities of machine learning and AI.

Project goals.

The primary objective of the Fake News Detection project in Liberia is to develop an automated system that can analyze news content, identify patterns, and assess the credibility of information, thereby enabling citizens to make more informed decisions. These are a few of the goals of this project: - Dataset Collection: Gather a diverse and representative dataset consisting of both legitimate news articles and examples of fake news prevalent in Liberia. This dataset will serve as the foundation for training and evaluating the AI models. - Model Development: Employ various machine learning techniques, such as natural language processing (NLP) and deep learning, to design and train models capable of distinguishing between authentic and fake news articles. The models will learn from the patterns and features present in the dataset to make accurate predictions. - Evaluation and Optimization: Evaluate the performance of the developed models using appropriate metrics such as precision, recall, and accuracy. Continuously refine and optimize the models to enhance their accuracy and effectiveness in detecting fake news. - User-Friendly Interface: Create an intuitive and user-friendly interface that allows users to input news articles and receive real-time feedback on the credibility of the information. The interface should provide clear indicators and explanations regarding the factors contributing to the classification of an article as authentic or fake.

Project plan.

  • Week 1

    Data Collection

  • Week 2

    Exploratory Data Analysis

  • Week 3

    Preprocessing

  • Week 4

    Feature Extraction

  • Week 5

    Model Development and Training

  • Week 6

    Model Evaluation

  • Week 7

    Model Integration

  • Week 8

    Model Deployment

Learning outcomes.

By participating in this project, participants will gain valuable knowledge and skills in the following areas:

1. Data Collection and Preparation: Acquire proficiency in sourcing and curating relevant datasets, ensuring their representativeness and quality for training AI models.

2. Machine Learning Techniques: Develop a comprehensive understanding of various machine learning algorithms and their application in natural language processing and text classification tasks.

3. Model Evaluation and Optimization: Learn how to assess the performance of AI models and employ optimization techniques to enhance their accuracy and reliability.

4. User Interface Design: Gain experience in designing user-friendly interfaces that enable individuals to interact with AI-powered systems effectively.

Demo video

https://liberia-fake-news-detection.streamlit.app/

Share project on: