How AI could make the internet a safer, more civil place | New York Post

kiratebbe's bookmarks 2018-03-15

Summary:

That’s why GLAAD announced this week it’s working with Google’s parent company Alphabet to change the way artificial intelligence understands LGBT-related content online.

“The internet is such a vital resource for the LGBT community, especially for young people finding connection,” he said. And that extends from YouTube to Google, Facebook and Twitter. “That lifeline is under attack.”

Because content related to marginalized or minority groups — gay people, women, people of color and some religions — tends to generate more negative feedback than content not related to those groups does, artificially intelligent algorithms have started to learn in some cases that LGBT-related phrases are “bad,” Halloran said.

To combat this, GLAAD is working with a division of Alphabet, called Jigsaw, to help the company train the artificial intelligence that controls online algorithms, teaching it which phrases are offensive to the LGBT community and which are acceptable.

Link:

https://nypost.com/2018/03/14/how-ai-could-make-the-internet-a-more-civil-place/

From feeds:

Harmful Speech » kiratebbe's bookmarks

Tags:

Date tagged:

03/15/2018, 15:20

Date published:

03/15/2018, 11:20