How Instagram Is Using AI As A Moral Compass For Abusive Comments

Instagram has invested heavily recently in AI helping to moderate and mitigate abusive and offensive comments and content. Now, Instagram's AI will attempt to act as your moral compass.

Think back to your childhood and you will likely remember instances where you were bullied and perhaps instances where you were the bully. These events are often imprinted in your memory, lasting long into adulthood.

The National Survey of Children’s Exposure to Violence found that 22% of students from age 12 to 18 said they were bullied at school. This trend has unfortunately not changed significantly over the past decade and with the increasing dominance of social media in youth lives the availability to bully while hiding behind a digital veil is increasing.

Credit: Monkey Business @

While the United States lags Europe in many regulations for the tech industry, there is increasing pressure on the social media behemoth Facebook to be a responsible player in data privacy and content moderation.

Instagram, a photo-sharing platform owned by Facebook, has taken to artificial intelligence as a form of content moderation, specifically targeting abusive comments and content and bullying.

While identifying and removing bullying on Instagram is important, we also need to empower our community to stand up to this kind of behavior, said Adam Mosseri, Head of Instagram in the announcement.

This provides insight into Mosseri’s train of thought on limiting bullying on the platform, it is a tricky balance of free and open social interactions and preventing abusive comments especially in youth.


The Current (Sad) State Of Online Bullying

A study came out last year from the Pew Research Center that outlined where cyberbullying is at today. What they found is that over half, 59% of teens in the United States have experienced online bullying. It is well documented that these bullying events can often lead to depression, negative thoughts, shame, humiliation, low self-esteem, and even suicide.

As adults, we realize the importance of combating online bullying and the prevalence that it exists today. Teens use social media as a way to express themselves, learn, and connect with the broader world. In doing so, they are put into a position where cyberbullying is a potential.

Conversely, nearly the same percentage of US teens that experience cyberbullying also do not tell their parents or guardians when or if they are victims of cyberbullying.

The Anti-Defamation League (ADL) published a fantastic list of warning signs parents can look for to get a sense of whether their children are being bullied online. These warning signs include:

  • A teen appears upset, irritable, sad or mad after online use
  • Overall signs of depression, including withdrawing from social activities, sleeping more than usual, lack of interest in things they used to enjoy, more emotional or reclusive than usual.
  • Unusual online activities that include changing the screen or app that they’re on when a parent walks by, a significant increase or decrease of online use.
  • Declining performance in school, activities, sports, etc. can all be signs of distraction caused by bullying.

There are, thankfully, many resources for parents or guardians of teens who are victims of cyberbullying. As discussed above, ADL is a great place to start for more information on cyberbullying.


AI-Powered Moral Compass – A New Feature To Limit Bullying On Instagram

The new feature, powered by AI is sort of the moral angel on your shoulder, urging you to reconsider before posting a hurtful comment.

Notification from Instagram’s AI on whether you really want to post the potentially harmful comment.

If you write what Instagram’s AI believes is an abusive comment it will add an intermediate step after you hit post. The comment will show up with the question “Are you sure you want to post this? Learn More.” The timer gives you a few seconds to undo the comment, in which case the user will never see your posted comment. Alternatively, you can tap learn more and it will pop up with a short description:

Keeping Instagram a Supportive Place

We’re asking people to rethink comments that seem similar to others that have been reported. If we made a mistake, let us know.

You can then tap through to the next step and provide feedback to Instagram if your comment was not, in your opinion, abusive. What I found is that the AI is inconsistent in its notifications. I posted the same comment above on other posts and never got the notification. Perhaps the AI algorithm is set to limit how many notifications people receive.

Given the notification that the comment appears similar to others that have been reported, the AI algorithm is clearly attempting to make links between previous comments people have flagged and new comments. The goal, stopping the comments before they are made in the first place.

In addition, Instagram has added a great feature to restrict cyber-bullies without them ever knowing. If someone consistently posts hurtful comments you can restrict them. This is instead of actually blocking the person or unfollowing them. If you restrict someone you will not see their comments nor will anyone else. However, the restricted person will see their comment, never realizing they are restricted. You can go in and manually approve comments from the restricted person if you want them to appear. The restricted person also won’t be able to see if you’ve read their DMs or if you’re active on Instagram.

Often times the person who is bullied may feel hesitant to confront the bullier and may know the person in real life. In this case, the restrict feature leaves the bullier in the dark and you don’t have to deal with their abusive comments.

Unfortunately, this feature hasn’t been launched yet so you will have to wait to restrict the bully. In the meantime, you can always follow these steps to report bullying on Instagram.

The yet to be released feature of restricting bullies on Instagram

What Can You Do Today To Stop Bullying On Instagram

What you can do today is to hide offensive comments automatically on Instagram. To enable this feature follow these steps.

  1. Open Instagram
  2. Tap your profile picture at the bottom right
  3. Tap the three horizontal lines at the top right
  4. Tap Settings
  5. Then Privacy
  6. Then Comments
  7. Toggle on “Hide Offensive Comments”
How To Hide Offensive Comments On Instagram

I wish this feature was a bit more obvious and easy to navigate to. 7 steps to stop bullying comments is too many in my opinion. When in doubt, report and block anyone who is posting abusive comments or bullying.

Protecting our youngest community members is crucial to helping them feel comfortable to express who they are and what they care about. – Kevin Systrom, Co-Founder & CEO of Instagram



How To Recover Deleted Facebook Messages

Your Facebook conversation is going great, you’re really hitting it off. Next thing you know, the other person it’s who you thought they were and you find yourself, in a […]

Titles For Friends: Fun & Cute Nicknames

No matter what your age is, we all rely on having best friends to lighten up our day, support us when we’re down, and go on epic adventures. Here we […]

Apps Like Offerup: Top 12 Sites To Sell Anything

While OfferUp is a fantastic site, there are alternatives that do various things differently and better. It’s just a matter of finding the right one that fits with your style […]

EARN IT Act encryption

The EARN IT Act: The End of P2P Encryption?

A new bill worming its way through Congress has drawn ire from free speech and digital privacy advocates. The new bill, introduced in early March and called the Eliminating Abusive […]

Android Messages App Is More Like iMessage Now Thanks To Google RCS Update

If you have an iPhone you know the difference between a blue and a green chat bubble. There is added functionality and more seamless messaging with iMessage (blue bubbles) as […]

How Many YouTube Channels Are There?

YouTube recently hit a huge milestone, over 1 billion hours of videos are watched each day. On top of that, the number of YouTube channels with over 1 million subscribers […]

What Can MMO’s Teach Us About Pandemics?

Pandemics by their very nature are difficult to predict. Human behavior is a major factor in the spread of disease and it’s extremely difficult to accurately predict human behavior in […]

How-To Apps is one of the fastest growing app websites in the world. We cover app technology, how-to guides, internet culture, and app news regularly. Since our launch just 8 months ago we've grown to help over 100k people per month on their app related questions. Want to know more?