SmartScrolling: An Innovative AI Approach for Detecting Misinformation on Social Media Platforms

Transcript:

Hi, my name is Emily Mineault and I am a grade nine student from Northern British Columbia. My project is called SmartScrolling: An Innovative AI Approach for Detecting Misinformation on Social Media Platforms. I developed SmartScrolling because I wanted to develop a way for misinformation to be detected online. SmartScrolling uses AI to help the app detect misinformation by searching the internet. I am working to turn it into a pop-up widget that appears when you are scrolling through social media that you can use to help you verify information. I did this because misinformation is everywhere online and many people have difficulty finding whether something posted on social media is real or fake. I also conducted research and surveys about social media and misinformation, trying to see how my app can help this growing issue. Check out the rest of my project to learn more! Thank you!

When I first began looking at project topics, I focused on issues that I saw were prominent and important in today’s world. Something I found to be incredibly concerning was the amount of misinformation I saw on social media, as well as how almost nobody ever seemed to point out that it was misinformation. I decided that for my science fair project, I was going to create something that could detect misinformation on social media. Misinformation detection and prevention matters to me because I think that it is incredibly important for people to not be misled online. Social media is different from regular news platforms in the sense that anybody could say anything online and label it as "real news," even if the information is false. I hope that my project can make the world a better place by helping people detect misinformation and debunk it.

After deciding on my project topic, I knew that I had to actually learn to code, so that I could develop my app. I decided to go straight into building SmartScrolling and to learn as I went, which I think was a great way to learn to code. There were many problems I ran into along the way, so I used Gemini AI as a tool to help me learn. Over the past few years, artificial intelligence has become a major component of coding, and I wanted to take advantage of everything that AI has to offer to the world of software development. Building the app was incredibly challenging; I knew very little about app development and I was learning as I went, being taught by Gemini AI, Thunkable, and my robotics teacher.

To develop my app, I used a platform called Thunkable. Thunkable uses the React Native language. There are a few reasons why I chose to use Thunkable instead of using other platforms.

1. It can be used on both IOS and Android.

2.It was fairly easy to learn to use.

3.It has Thunkable AI incorporated into the development process, which was very helpful to me with fixing bugs and issues.

For my backend, I used Firebase, as I found that it was recommended lots online.  It wasn’t the easiest to use, as there was a learning curve, however I eventually figured it out. I used the Blaze plan on Firebase, which worked well for a project like this.

For my API keys, I used my OpenAI API key to access the OpenAI Whisper model. I used this Whisper model to transcribe the TikTok videos. I also used a Gemini API key, which worked to analyze the videos and detect the misinformation.

Here are my survey results! I surveyed 16 people of varying ages.

As we can see from the surveys, 81.3% of people surveyed said that they believe an app like SmartScrolling would be either helpful or extremely helpful. 68.8% of the participants surveyed stated that social media is where they receive the most information. 50.1% of people surveyed said that they are either confident or very confident detecting misinformation on social media. This shows that only about half of the people who were surveyed are confident in their ability to notice misinformation on social media. The participants were also asked how often they do research regarding something they saw on social media, and only 18.8% of people said that they always research further to confirm the information.

SmartScrolling is still in the development process at this time (April 30th, 2026), however the prototype is working fairly well. It is able to read text and send back an analysis of the text. However, I am still working on how it grabs links to social media posts.

As shown throughout the survey, we can see that social media is a major news outlet for many people. However, there are few people who consistently fact-check information when scrolling on social media platforms. I believe that SmartScrolling is an important first step to preventing the spreading of misinformation on social media. Looking forward, I would like to work on publishing SmartScrolling to be available for users, and I would also like to grow SmartScrolling as a company.

I would firstly like to thank my delegates, Mr. M. O'Brien and Mrs. J. Copeland. I also give a huge thank you to my robotics teacher, Mr. W. McColm, for teaching me how to code! I would also like to thank my parents for being my biggest fans and my loudest supporters!

Allen, S. (2023, March 31). How Social Media Rewards Misinformation | Yale Insights. Yale Insights. Retrieved April 6, 2026, from https://insights.som.yale.edu/insights/how-social-media-rewards-misinformation

American Psychological Association. (2023, November 29). How and why does misinformation spread? American Psychological Association. Retrieved April 6, 2026, from https://www.apa.org/topics/journalism-facts/how-why-misinformation-spreads

Government of Canada. (2023, December 20). Concerns with misinformation online. The Daily — Concerns with misinformation online, 2023. Retrieved April 6, 2026, from https://www150.statcan.gc.ca/n1/daily-quotidien/231220/dq231220b-eng.htm

Micich, A., & Cross, R. J. (2025, July 30). How misinformation on social media has changed news. PIRG. Retrieved April 6, 2026, from https://pirg.org/edfund/articles/misinformation-on-social-media/

SWGFL. (n.d.). Misinformation on Social Media – Guidance, Impact and Support. SWGfL. Retrieved April 6, 2026, from https://swgfl.org.uk/topics/social-media/misinformation-on-social-media-guidance-impact-and-support/

Google. (2026). Gemini. (February and March 2026 versions) [Large language model]. https://gemini.google.com/app