Schools Are Using AI to Check Students' Social Networks for Warning Signs of Violence

[froggyboy604]

In the hopes of deterring violence, schools are turning to big data analytics to examine social media posts for the earliest signs of violence—depression, resentment, and isolation. Shawsheen Valley Technical High School in Massachusetts has turned to Social Sentinel, a data analytics company that says it can use the type of threat detection police agencies use to identify students at risk. But experts worry student social media mining, even with the best intentions, is a slippery slope to treating students the way we treat suspects.

https://gizmodo.com/schools-are-using-ai-to-check-students-social-media-for-1824002976

I feel spending more time on recommending mentally ill student go see a therapist or in more serious cases spend time at a mental hospital to get more treatment can prevent more violence at school.

Most likely using data from social networks to prevent student violence can lead to a lot of posts where non-violent students post violent short films to YouTube, or more violent rap lyrics to Facebook and Twitter for fun.

 

[テクニカル諏訪子]

Last time I was a system administrator for a large school, I had to check the logs for similar things every single hour by hand.
Luckily I'm no longer need to do this any more, though I'm very skeptical about the AI.

At least me as a human being can verify whether the use of SNS during school time is justified or not, an AI is even prone to all sorts of false positives, or even false negatives.
And AI can easily be manipulated for propaganda purposes too like censoring actual information and favouring false information, which is what has been discovered on sites like YouTube, Facebook, Google, and especially Twitter long ago.

 

[froggyboy604]

I think AI may also not be fast enough at warning which students post threats to the school. It maybe faster for other social network users, students, teachers, and tech staff who report threats to the police or principal when they see a violent threat.

The huge number of false positives and negatives may cause staff to ignore the warning e-mails or make it difficult to find real violent threats against a school because the logs are very long to read for a human, and a lot of kids may post a lot of fictional posts on violent movies, music, games, and books which may cause the AI to be overloaded with false positives.

 

[Demon_Skeith]

Bad time to be talking about violent video games like shooters and such. This is why you stick to places like forums where you can use a username of your choice. Real name online is always bad unless in a professional place like linkedin

 

[froggyboy604]

Bad time to be talking about violent video games like shooters and such. This is why you stick to places like forums where you can use a username of your choice. Real name online is always bad unless in a professional place like linkedin

I agree forums are the best place to talk about violent video games because you can use a username of your choice.

I think it is okay for talking about violent video games with a patriotic US Army theme like America's Army on social networks. Talking about patriotic US Army games can make you look more patriotic in places with a large US Army base.

Anime/Cartoon based video games with violence like Kingdom Hearts are still okay to talk about since these video games have been around for years without much complaints about violence.

But, talking about crime theme games like Grand Theft Auto can be a bad time on social networks which require users use a real name.

 

[Claraviolet]

This is the best thing that I have read after a very long time.
May be they could even detect bullying and eradicate it completely.