This week the hugely popular social networking platform Instagram, announced the launch of its new features designed to tackle messages which might be considered as bullying or harassment.
According to an article on the company’s blog ‘We can do more to prevent bullying from happening on Instagram, and we can do more to empower the targets of bullying to stand up for themselves’.
Adam Mosseri (Head of Instagram) says ‘ The intervention gives people a chance to reflect and undo their comment and prevents the recipient from receiving the harmful comment notification’.
This sounds great. Instagram are right. They badly need to focus on abusive posts. The new feature tool encouraging the opportunity for reflection and the chance to undo is very welcome. We’ve all texted posted on impulse. And regretted it later. The statement is however misleading in that it will not always ‘prevent the recipient from receiving the harmful comment notification’. Instagram did however admit that at test it worked with ‘some people’. We’ll take that. Some is better than none and it’s time reflection tools were introduced.
We’re big fans of social media and social media companies taking more responsibility for what happens on their platforms so were pleased to see Instagram making efforts in this area. While bullying is an age old problem offline, online it happens predominately via social networks and gaming so it’s vital the companies make efforts here. Let’s face it, there are very few social media platforms where some form of bullying, harassment or abusive behavior doesn’t take place.
The first feature being rolled out claims to be ‘AI that notifies people when their comment may be considered offensive before it’s posted’ and supposedly detects ‘bullying and other types of harmful content in comments, photos and videos’.
Given that there are many different variations of words and also messages contained in images and videos, sometimes this can be difficult to pick up or detect.
Having worked on a previous project along content identification on a huge scale, using algorithms, I know from experience this is something which is difficult to get correct/right a lot of the time.
Does it stop it, make the poster think or just give a subtle prompt?
The second feature called ‘Restrict’ is not something which is innovative or new. In fact this feature has been widely available for years (shadow banning) though again it’s good to see it being used on a prolific platform like Insta.
The idea is that the person targeted with the abusive posts can prevent them from being seen publicly. Their messages are purged elsewhere. The main reason for this feature is that the poster of abusive comments still thinks they are posting publicly. When they get no engagement the idea is that they will hopefully get bored/stop.
In a nutshell the ‘Restrict’ feature works along the same principal according to Instagram ‘Once you Restrict someone, comments on your posts from that person will only be visible to that person. You can choose to make a restricted person’s comments visible to others by approving their comments’.
Anyone involved in any type of mental health or psychological sphere will tell you that developing technology around these types of harassing or bullying behaviours will take time and we’ll not see the impact for some time and only then with a complex combination of both human processes and machine learning. The examples used by Instagram on posts use very basic language eg ‘You are so ugly’. In real online life abusive language is a lot more nuanced and therefore harder for AI to detect. Nonetheless, they are steps in the right direction.
Remember to share it with your family & friends.
Business, Career & Life Podcast
