Facebook is known to use artificial intelligence (AI) throughout its features. Instagram as a company that has been acquired by Facebook, it should too use AI as a next logical step up from previous attempts.
Social media in overall, is a cruel place to be. Becoming users of the network comes with risks of privacy, hate speech and offensive heartless comments.
Instagram starts using AI in a way to combat hate speech that also common within its platform.
In 2016, Instagram has introduced a keyword filter that allowed users to automatically remove specific offensive words that appear on their feed. What tool did, was removing words that are "often reported as inappropriate", and let users set custom keywords.
But this tool was good only for a while. Fighting hateful comments can be exhausting, and it has exhausted the tool as internet trolls were able to come up with new offensive languages every time they were caught.
From misspelled words to slurs, users on the internet can even create entirely new words just to offend people.
To fight trolls, Instagram has to do more than filtering words. Using AI, Instagram uses filters that learn to understand the context around offensive speech. With machine learning, the AI should be able to rid abuses even if the purposely offensive word doesn't trigger Instagram's existing filtered keywords.
This is Instagram's hope to mitigate the endless fight against hateful comments that irritated many users.
"Many of you have told us that toxic comments discourage you from enjoying Instagram and expressing yourself freely," wrote Instagram CEO and co-founder Kevin Systrom in a blog post. "To help, we've developed a filter that will block certain offensive comments on posts and in live video."
The AI is called DeepText, an artificial intelligence effort developed in-house at Facebook. What it does, is to identify offensive languages. Using machine learning, the system should get smarter over time as it learns different meanings in different context.
Besides putting it to filter hate speech and offensive language, DeepText is also being used to identify spam.
To do what it is supposed to do, DeepText was trained to identify negative comments and categorizing them into broad segments like "bullying, racism, or sexual harassment". The raters are said to have analyzed at least two million comments before the system went live.
Instagram's DeepText system follows Facebook's success in fighting spam. After training the system with human input on how to ID spam, the team liked the results. But they didn't exactly say how much DeepText's implementation was able to reduce spam versus previous methods.
The AI is meant to work for the good cause. But if users don't like the idea of a machine deciding what comment is good or not, they have the option to turn the feature off from the comments section at Instagram's Settings.
