Fake news detection within online social media using supervised artificial intelligence algorithms
Did you know that media tracking algorithms can be trained across both social networks and news agencies to search for tell-tale indicators that some maybe not factual at all?
Belgium-based startup VeriFlix has developed a fascinating application that uses this kind of technology. They have developed a method of scanning user-submitted videos, which play an increasingly important role in most media organisations’ performance and determine if the content is based on factual information.
The company’s technology is now being used by one of the largest media outlets in that region, including Rourlarta, after winning funding through Google’s Digital News Initiative, with promising results.
This application initially was a peer-to-peer video crowdsourcing app, then kept on evolving into the idea where media organisations allow any user to start recording and submit videos with the correct geolocation and metadata.
The videos get sent to the platform where VeriFlix adds a layer that first detects every stream’s content — so they can say what we see in the video, alongside the geolocation data and time stamp.
Once the videos are tagged, VeriFlix can compare them with other videos. An example would be that one request results in 1,000 videos. VeriFlix then can compare the content of every video. Suppose a majority of the videos show the same content. In that case, it can verify the authenticity of what has been shot.
VeriFlix uses the real-time object detection algorithms YOLO (You Only Look Once) to identify and mark video content before moving the data on to proprietary algorithms built in collaboration with KU Leuven University. These algorithms analyse the data, and timestamp and geolocation data passed through the application’s safe interface.
There are various advantages to this kind of technology, one being that videos are verified to be authentic as they are shot in the location where they say they are taken, and not modified or doctored, another being that it is able to bridge the gap between media houses and the audience — letting the audience be part of the story.
VeriFlex stated that this technology will not replace the journalist’s job but can help journalists get the raw data.
Bibliography
Degeler, A. (2018, June 12). Funded by Google’s journalism grant, VeriFlix takes on validating video content from citizen reporters. Tech.Eu. https://tech.eu/brief/funded-googles-journalism-grant-veriflix-takes-validating-video-content-citizen-reporters/
Heathman, A. (2018, June 12). This new AI technology wants to help journalists fight fake news. Evening Standard. https://www.standard.co.uk/tech/google-funded-ai-technology-fake-news-a3860611.html
Marr, B. (n.d.). This Google-Funded Company Uses Artificial Intelligence To Fight Against Fake News. Forbes. Retrieved December 11, 2020, from https://www.forbes.com/sites/bernardmarr/2018/07/18/this-google-funded-company-uses-artificial-intelligence-to-fight-against-fake-news/#17a0523ca4b5
This Google-Funded Company Uses Artificial Intelligence To Fight Against Fake News. (n.d.). Bernard Marr. Retrieved December 11, 2020, from https://bernardmarr.com/default.asp?contentID=1538
Using AI to fight fake news. (n.d.). Broadcast. Retrieved December 11, 2020, from https://www.broadcastnow.co.uk/tech/using-ai-to-fight-fake-news/5131397.article
This blog is a project for Study Unit MCS5460, University of Malta.
