How the 5 Biggest Social Media Sites Use AI

There is a good chance that you or someone you know uses social media

Over half of the world’s population uses at least one social media website. Whether through Facebook, Twitter, Youtube, etc., social media provides an opportunity to connect with others more efficiently than ever. 

Social media platforms need technology that keeps up with massive amounts of users. That’s where artificial intelligence comes in. With AI, social media platforms can keep up with monitoring their platform and curating content that users are most likely to enjoy and stay on the platform.

AI has become an essential aspect of social media sites in many ways, which is why many software engineers can make up to six figures a year.  

Here is how the five most prominent social media sites implement AI to improve their platform: 

Youtube

With over 500 hours of video footage uploaded to Youtube every minute, it may prove challenging for employees to regulate the massive amount of content. Google (Youtube’s parent company) has introduced a new “trashy video classifier” system entirely operated by AI. 

Youtube has faced controversies with troubling content uploaded to the platform. One controversy, known as “Elsagate”, involved a plethora of inappropriate videos disguised as being “child-friendly” uploaded to the Youtube Kids website.

Google’s AI system is capable of:

  • Scanning Youtube’s homepage for troubling content.
  • Analyze videos for misleading titles/content.
  • Review reported videos by users.

Youtube’s video analysis technology is essential to manage the workload of monitoring content and draw advertisers to the website.

Whatsapp

In an age where information is rapidly available, disinformation finds its way into various social media platforms. One of the most affected is the instant-messaging application, Whatsapp.

In 2018, around a dozen mob lynchings occured in India fueled by rumors circulating around Whatsapp involving child-trafficking and organ harvesting. These rumors were simply not true and innocent people’s lives were lost. 

A Spanish fact-checking organization, Maldita, created an AI chatbot to combat this quick spread of false information on its platform. 

When a user sends circulating information to Maldita’s Whatsapp number, the automated bot will analyze the content and provide the previous fact-checked articles on the information. Or the bot will notify journalists if what is sent is not fact-checked. 

The AI bot is designed to regulate “fake news” by automatically analyzing content that coincides with Maldita’s previous fact-checks. Bots like these help platforms control false content that could potentially become dangerous

Facebook

With 2.9 billion monthly users, Facebook has become a giant in social media platforms. Meta (formally known as Facebook) has been shown to push boundaries with AI implementation to run a successful website. 

One way Meta uses AI to understand its Facebook users is through Deeptext, a deep-learning text-based engine.

Deeptext can:

  • Understand textual contexts of posts with “near-human” accuracy
  • Analyze around 20,000 posts per second
  • Pick up on slang words and understand 20+ languages

By converting words into a computer algorithm, Deeptext can understand the similarities of words based on the integers assigned to them. Deeptext enhances the Facebook users’ experience by comprehending their posts and, based on that, provides desirable content and product suggestions for them. 

Twitter

If you want to experience opinions and news from around the world, Twitter is the social media platform to visit. 

Twitter engineers now use deep learning technology to improve their algorithms and provide content catered explicitly to users

The technology considers many factors when it designs a user’s timeline. This includes the tweet’s overall engagement, the author of the tweet and their relationship with the user, and the type of tweets the user has engaged with in the past. 

Although the concept seems relatively simple on the surface, the intricacies of designing the unique timelines of 330 million monthly users can realistically only be achieved using AI software to reduce workload and provide efficiency.

Instagram

Instagram is a visual social media platform with users typically portraying messages through images only. Also, users can create short videos like Reels and stories by utilizing various tools like Instagram story makers.

To enhance the visitor’s experience, Instagram implements machine learning by analyzing data around posts and determining whether to include them in their algorithm.

By using insights (or “signals”) of a post, the ML algorithm ranks the publications according to: 

  • The popularity of a post
  • How many times a specific user has interacted with another user
  • How much time an Instagram user spends on different pages
  • History of a post’s interaction 

By analyzing these “signals,” Instagram can keep users happy by providing content people want to see.

Contributor

Cameron is pursuing his BA in journalism at the University of Houston, Texas. Before pursuing his degree, his fascination with content creation and the evolution of technology has stuck with him since a young age. His goal in life is to write and create media to showcase technology’s development and impact on society.

About Cameron Schoppa

Cameron is pursuing his BA in journalism at the University of Houston, Texas. Before pursuing his degree, his fascination with content creation and the evolution of technology has stuck with him since a young age. His goal in life is to write and create media to showcase technology’s development and impact on society.

View all posts by Cameron Schoppa →