Facebook Artificial Intelligence Research: How the Social Media Giant is Using AI

While big tech companies — like Alphabet, Microsoft, Apple and Facebook — are all known for different products, they are united in one thing: pushing the boundaries of AI technology with new research.

Part of the driver behind research in artificial intelligence has been the growing amount of data these companies collect. When you collect information at a large enough scale, it can become too much for human researchers or conventional analytical techniques. With AI, it’s possible to tackle these massive data sets and find extremely subtle patterns and relationships. The information that an AI can uncover can easily reshape everything from content recommendations, to marketing, to social media moderation. 

For years, Facebook has had one of the most aggressive data collection strategies in place. Now, Facebook is one of the tech companies at the forefront of AI research.

What Is Facebook Artificial Intelligence Research (FAIR)?

FAIR is Facebook’s in-house AI research lab. According to the lab’s website, their researchers are currently investigating a wide range of topics, including conversational AI, computer vision, natural language processing and data integrity technology. 

The lab has been around for a little less than seven years,and has had a big impact on both Facebook’s operations and the AI research community during that time. 

Deep learning — an AI approach where artificial intelligence “learns” from new data and upgrades its pattern-finding algorithm over time — is central to Facebook’s operations. “You wouldn’t be able to run Facebook without deep learning,“ Facebook Chief AI Scientist Yann LeCun told TechCrunch in 2018.

The FAIR research lab, unlike Facebook’s more product-focused Machine Learning team, works on more long-term projects, giving the team extra freedom to experiment and push the boundaries of AI’s capabilities.

For example, the lab’s researchers have tackled a number of computer vision projects, which aim to create algorithms that help computers “see” the world around them. A highly reliable version of these algorithms will be necessary for tech like safe self-driving cars, or autonomous drones and robots.

What Facebook Is Researching Right Now

Facebook continues to make major strides in AI — with the company’s lab making several major announcements even just this year.  

One example is a new natural language processing AI that can translate between more than 100 hundred languages, and could help facilitate international communication on the platform. Like other FAIR projects, the algorithm is being open-sourced to the AI community, meaning researchers from outside Facebook will have full access to the AI’s code.

Challenges for Facebook’s AI Research

However, despite the advances in AI technology Facebook has made, the company’s use of AI hasn’t been without difficulties.

In the early days of the COVID-19 pandemic, social media platforms struggled with misleading posts that ranged in content from outdated research to outright misinformation. 

Facebook’s content moderation and recommendation AIs had proven successful at identifying and removing harmful content like hate speech. However, Despite the company’s best efforts, its AI models weren’t able to identify disinformation, and human moderators had to step in to flag potentially misleading posts.

The failure comes as social media companies like Facebook turn to AI algorithms to moderate their platforms. Challenges like these show how trusting AI to moderate massive social media sites could be a mistake in the long run.

The Future of AI and Facebook’s Artificial Intelligence Research

Facebook, along with other Big Tech companies, will likely continue to be a driving force behind AI research. The massive amount of data the platform has access to means that researchers will have the best available information and data sets for working on AI projects — and also that AI will be essential to the future of Facebook’s success.