[ad_1]
About six Nina last year, Nina Nørgaard met weekly for an hour with seven people to discuss the sex and violence used against women on television. Nørgaard, a PhD student at IT University of Copenhagen, and his discussion team were making extraordinary efforts to identify misogyny online. The researchers paid the seven to look at thousands of Facebook, Reddit, and Twitter and see if they were displaying immorality, misconceptions, or harassment. Once a week, the researchers brought in a team, with Nørgaard as mediator, to discuss issues they did not dispute.
Misogyny is a pandemic that affects how women are represented online. A 2020 Plan International learning, one of the largest of its kind, found that more than half of women in 22 countries said they had been abused online. One in five women who have been sexually assaulted said they had changed their behavior – reducing or stopping using the internet – because.
Television companies use it artificial intelligence recognizing and removing records that denigrate, abuse, or threaten violence against women, but is a serious problem. Among researchers, there is no level of exposure to sexually explicit material; a recent paper identified four categories of complex issues, while the other identified 23 categories. Most studies are in English, leaving people working in other languages and cultures without complex guidance and often their choices.
That’s why researchers in Denmark tried a new method, writing Nørgaard and the seven people regularly to review and write notes, instead of relying on full-time contractors. paid by post. He deliberately chose people of different ages and nationalities, with different political views, to reduce the likelihood of bias toward the same cause. The staff included a programmer, weather forecaster, actor, and health worker. Nørgaard’s work had brought them to unity.
“The interesting thing is that they don’t agree. We don’t want a vision. We don’t want everyone to think the same way, ”says Nørgaard. He said his goal was to “bring them into contact with each other or in public.”
Nørgaard felt that his job was to help the staff “find the answers.” Over time, they became acquainted with each of the seven individuals, and who, for example, spoke more than others. He tried to make sure that no one was leading the discussion, because it should be a discussion, not a discussion.
The most difficult singing was to record with jokes, jokes, or insults; became the main topics of discussion. Over time, “the meetings became shorter and people didn’t talk much, which is why I see it as a good thing,” says Nørgaard.
Researchers who do this work call it success. He says the discussions led to the development of written information to support AI aligorivimu. Researchers say that AI that is well integrated with the content can detect misogyny on well-known platforms at 85% of the time. One year earlier, the most common misogyny detection method was accurate about 75% of the time. Overall, the group has also reviewed some 30,000, 7,500 articles that are considered cruel.
The articles were written in Danish, but the researchers say their method can be used in any language. “I think that if you can explain singleness, you have to follow a path that has a lot of our stuff. Otherwise, you risk cheap information, and it undermines everything,” said Leon Derczynski, co-author and fellow professor at IT University of Copenhagen.
These findings may be useful beyond the media. Entrepreneurs are beginning to use AI to display work lists or public documents such as sex journalists. If women isolate themselves from online discussions to avoid abuse, this will undermine democracy.
“If you do not look at the threats and violence of half the population, then you will not have the best opportunities for online democracy as you can,” Derczynski said.
[ad_2]
Source link
