Science Google Developing Troll-Detection Software

http://www.theverge.com/2017/2/23/1...ctive-software-ai-machine-learning-developers

Google’s Jigsaw unit, as part of a larger effort to battle online trolling, said earlier today that it was releasing a new tool called Perspective, software that uses machine learning to detect harassment and abuse online.

While Jigsaw and its software are not new — it’s been around in some form since 2010, though with a different name, and Jigsaw software has even been used to spot potential ISIS recruits — this new Perspective API makes it available to developers who want to use and build on top of the software. The software works by determining the “toxicity” of online comments, a scale that has been established by mining millions of comments from the web and then presenting them to panels of 10 people (humans!) at a clip to get their feedback.

Try using Perspective to determine the toxicity of your average post

http://www.perspectiveapi.com/

@Null should embed this to grade our posts based on how toxic they are

Spoiler

 

[alex_theman]

Someone needs to create a bot for this. So we can make it not work.

 

[Null]

How has this not been tried yet?

View attachment 186195

omfg you fucking cunt you stole it from me

Spoiler

 

[Kari Kamiya]

Just dicking around for now but

Spoiler: that escalated quickly

On the other hand:

What did Sonichu ever do to you?

EDIT:

Every little thing counts.

 

[ICametoLurk]

First one to get one hundred percent troll gets all my good points.

 

[Sperglord Dante]

Sobchak is a problematic word.

 

[Ruin]

Holy fuck.

 

[Doseferatu]

Apparently, raping someone and not giving them HIV is worse than trying to give them HIV.

 

[Kari Kamiya]

Name test.

Spoiler: Priorities

I also accidentally spelled the wrong "Joseph" and

 

[HG 400]

 

[AnOminous]

Very good! I was theorizing it would be a short one and trying everything I could think of.

 

[Dandelo]

 

[Techpriest]

Yeah this software sucks.

 

[AnOminous]

Got it down to three words.

 

[muina]

 

[Kari Kamiya]

get rekt tumblr--

--fuck.

 

[GethN7]

This was surprisingly toxic:

I tested it with some other phrases

"You can die" (81% similar to comments people said were "toxic")

Odd, this is a statement of fact.

"Death" (71% similar to comments people said were "toxic")

That's the word alone, devoid of context.

 

[Techpriest]

The real problem with this is the fact it's taking feedback from the public to influence the algorithm. "Hitler did nothing wrong" can be untoxic'd based on the feedback. This a good way to totally fuck up this project. Tell the bot that 'Hitler did nothing wrong' isn't toxic, or whatever else you want, and this thing becomes useless pretty quick

 

[AnOminous]

Congratulations to Google for creating the world's first shitposting practice tool.

 

[GethN7]

Okay, Google has some explaining to do:

 

[Kari Kamiya]

A great number of paleontologists took this too personally.