Grants:IdeaLab/Web of Trust + AI

Inspire campaign

Web of Trust + AI

Build a web of trust on top of Wikipedia. There should be different levels. For example, if you know someone personally and can vouch for them this should be the highest level. Penalize not just the violator, but also those that trusted him or her if there is harassment. In addition to this, use machine learning to gauge whether a message or edit is potentially a "hot" word ("bitch", "whore", etc). Use standard machine learning techniques (ROC curves, Factor Analysis / PCA, etc) to achieve ever increasing accuracy in determining whether a comment is harassing. When unsure, use a volunteer to assess. Make the data sets open so that everyone can expand on the techniques used.

idea creator• Zachaysan

this project needs...

volunteer

developer

designer

project manager

community organizer

researcher

created on19:05, 3 June 2016 (UTC)

Project idea

What is the problem you're trying to solve?

I'd like to stop users from getting harassed on the internet.

What is your solution?

Build a web of trust on top of Wikipedia. There should be different levels. For example, if you know someone personally and can vouch for them this should be the highest level. Penalize not just the violator, but also those that trusted him or her if there is harassment. In addition to this, use machine learning to gauge whether a message or edit is potentially a "hot" word ("bitch", "whore", etc). Use standard machine learning techniques (ROC curves, Factor Analysis / PCA, etc) to achieve ever increasing accuracy in determining whether a comment is harassing. When unsure, use a volunteer to assess. Make the data sets open so that everyone can expand on the techniques used.

Get Involved

About the idea creator

I'm a Software developer from Canada with experience in Linux based web development and machine learning.

Endorsements

• There is another idea like this but this one looks more professional. Victorvic1 (talk) 20:55, 13 June 2016 (UTC)

Expand your idea

Would a grant from the Wikimedia Foundation help make your idea happen? You can expand this idea into a grant proposal.

Expand into a Rapid Grant

Expand into a Project Grant
(launching July 1st)

Similar Ideas

Maybe we should merge this ideas.

• Machine-Learning to analyze and block content with harassment
• Use Facebook's DeepText to classify if a new post has potential harassment
• Anti-harassment bot