Combating Misinformation with Data Science and AI

With 91% of Nepali Population having access to the internet and 65 % users on mobile internet users, digital media and social networks are key for spread of information. The ability to follow a group, share information, react, comment and re-share any post have made a common person a social media star in a matter of days. Also the same media can damage someone's reputation, build distrust towards media, politics, established institutions and governments very quickly. In fact, Fake news spreads like wildfire faster and farther than true stories, and humans are primarily responsible for the spread of misleading information.

More so, during local and general elections in Nepal, we came across a flood of potentially false claims in the media, many of which we assume to be true. There are many cases we can see where political parties misuse social media platforms during elections to advance their populist agendas.

Not everything that one sees on the internet can be believed, yet when we are browsing, we generally don’t seek the source of the information.

There are different Types of Misinformation:

• ClickBait.
• Propaganda.
• Sponsored contents.
• Satire and hoax.
• Misinformation.
• False news.
• Disinformation.
• Rumours.
• DeepFakes and Manipulated photos and video.
• Posts designed to type RIP and comment to get viral.

Different types of problems require different types of solutions.

Though there are many fact checking organisations (eg: NepalCheck.Org , Nepal Factcheck, South Asia Check, Media Action Nepal) many fake news still spread in the social media. The goal of the project is to provide a platform to allow individuals to check if a news content/ post is already fact checked and identify it as fact or fake.

• Source through different Nepali fact check organisation to collect a dataset of verified fake news.
• Carryout data processing and data analytics to understand the distribution of misinformation.
• Apply Data Science, Data Engineering and Machine Learning develop an API to identify if a particular post has already been classified as misinformation.
• Develop a front-end and host a platform to check if a post has already been classified as misinformation.

1. Data Collection: Source and scrape news and post on fake information.
2. Data Cleaning.
3. Data Analysis.
4. Building Machine Learning Models.
5. Developing an API.
6. Building and hosting the platform.