Facebook unveiled a natural language processing engine Wednesday in its quest to create a deep learning-based artificial intelligence that can understand how humans talk with a near-human level accuracy.

The new AI, called DeepText, is designed to understand what kind of content Facebook users want to see, and it is capable of comprehending the textual content of several thousand posts per second across 20 languages.

Telling the difference between certain nuances in language has been a major issue for researchers trying to create truly human AI, and Facebook hopes to solve this problem with DeepText by understanding the words individually and their context in order to comprehend their intent.

"Traditional techniques require extensive preprocessing logic built on intricate engineering and language knowledge. There are also variations within each language, as people use slang and different spellings to communicate the same idea," the social networking giant explained in a blog post. "Using deep learning, we can reduce the reliance on language-dependent knowledge, as the system can learn from text with no or little preprocessing. This helps us span multiple languages quickly, with minimal engineering effort."

DeepText, as a result, can understand the difference between messages like "I need a ride" and "I found a ride" to see whether or not you need an Uber. It could also automatically format posts with price and item details if it detects that you're trying to sell something in a status update, as well as show you kinds of comments on celebrities' posts that you may actually want to read.

Another benefit of Facebook's new AI is that it would be able to stop spam from getting into filters and your feed by understanding what these posts say and look like.

The effect that DeepText would have on privacy has been a major concern among Facebook users, as the company noted that the AI scans the content of users' private messages, which previously resulted in a class action lawsuit for the site. However, TechCrunch notes that this could help DeepText pick out posts involving hate speech and harassment.

Being better able to route content from authors to viewers could help Facebook stay ahead of Twitter, SnapChat and other popular social networks, as well as other tech giants that are working on their own deep learning AIs, such as Google and its "SyntaxNet" system.