Can artificial intelligence really stop bullying online?

News dropped yesterday that Instagram is going to attempt to lead the fight against online bullying by “encouraging positive interactions”.

What was interesting wasn’t so much the AI that’s helping to power the initiative, but more the way in which one part of the plan places much greater emphasis on the ‘troll’ (for want of a better term), rather than the person being trolled or bullied.

Adam Mossieri, Head of Instagram, explained that the feature “notifies people when their comment may be considered offensive before it’s posted.” Consequently, Mossieri suggests, this is supposed to “[give] people a chance to reflect and undo their comment and prevents the recipient from receiving the harmful comment notification.”

This is a novel approach. To a certain extent, it’s probably a smart idea for an app to try and restore some semblance of self-awareness to users that have, up until now, been lured into a sweet ecstasy of evocative landscapes and thirst traps.

The drawbacks

However, a psychologist talking to HuffPost UK suggests that the feature might not actually be that effective.

Philip Karhassan likes the idea of a time out mechanism for users, but remains sceptical about its ability to actually change behaviours.

Such is the nature of how we have been trained as users to respond to in-app messages and notifications, the noble do you really want to say this bro? message will most likely be eventually ignored.

“Your brain just automatically finds a new way of relating to that pop-up,” says Karhassan.

The forthcoming ‘Restrict’ feature

This is only one part of the plan. Instagram also wants to “empower” victims of bullying on the platform using a feature they’ve called ‘Restrict’.

“Once you Restrict someone, comments on your posts from that person will only be visible to that person,” Mossieri explains. “You can choose to make a restricted person’s comments visible to others by approving their comments.”

Can AI keep up with the changing contexts and expressions of hatred?

There’s probably a lot here that other platforms, like, say Twitter, can learn from.

But while it’s good that users will have more power and control over who is able to interact with them, if we think more broadly about the ways in which platform capitalism facilitates hate speech and - potentially - violence, we’re not strictly dealing with disparate issues of personal civility.

Ultimately, there needs to be a sustained conversation about what these platforms are for and how we want to interact with one another. Safer spaces are an important start but whether the artificial intelligence that’s supposed to restore control to users actually dealing with abuse remains a whole other issue.