Detect and stop hate speech, harassment and cyberbullying

Started by
4 comments, last by ongamex92 7 years ago
We have developed a product to help online games detect and stop hate speech, harassment and cyberbullying, creating toxic free discussions and environments.

If you need something like this in your game let us know. Check our product here: https://sherloq.io
Advertisement

https://www.dropbox.com/s/y61q88pzgo293pe/Oops.png?dl=0

So my first test of your system had it fail. This would qualify as hate speech honestly. I by no means believe this, I was merely testing the system. It did catch a few individual swear words mind you. But these were the first two sentences I wrote that I could think of that might skirt a system like this.

This is a system I might be interested in for our next game if it's priced reasonably. Best of luck!

In case I delete the image in the future, the text is:
"I'm a big fan of hitler and all he did in life."
"I believe his final solution was the right idea in all respects."

I'll give the system the benefit of the doubt on the second sentence, as it requires the first to make sense honestly. But the first sentence should have been disallowed in my opinion. Having run and played on tons of online games the number of times hitler is invoked to get a rise out of someone else is not easily counted :)

"Those who would give up essential liberty to purchase a little temporary safety deserve neither liberty nor safety." --Benjamin Franklin

That's a good idea for a product. Maintaining a healthy online community is a problem the larger a game gets...

What kind of business model are you planning to use?

It's going to be pretty hard to get it to learn sarcasm though. The first sentence here is toxic as "tards" is an insult - your system should be able to learn that pretty easily. The two other sentances are technically polite things to say to someone, but in this context they're actually insults due to the sarcasm :o

"You do know the zealot rush is for tards, right? Nice one, buddy. You deserve to win with that strategy."

This text is healthy
Hmm. "Tits or GTFO" was deemed healthy. I think the system still needs some work! I like the concept, though.

Hmm I greatly approve of this - slightly sarcastic. Its a good concept, right now though basically all classic "British" insults get deemed as healthy as im guessing the creators arent from the UK. Heh, means us brits can still freely insult the world but they cant insult back ;p

Edit: Tried a few legitimate non-hate based sentences - explaining definitions of things etc and some that contain a word that it seems to flag and they tended to come back as ones it would block.

And some false negatives
"Being straight is wrong." - "This text is healthy"

"Faggot is a bad word, don't use it." - "This text is offensive"

"Fuck the police." - Says it is healthy :P

EDIT: Now that I think about it these may be a bit unrealistic in a game scenario.

This topic is closed to new replies.

Advertisement