Next month,Tumblr will start cracking down on "non-consensual sexual images," including deepfakes and creepshots,as part of an update to its community guidelines,reportsThe Daily Dot.The microblogging platform—are we still calling it that???—is also introducing tighter restrictions on hate speech and violent content.
"Our Community Guidelines need to reflect the reality of the internet and social media today and acknowledge that the things people post and share online influence the way others think and behave," explains a blog postannouncing the changes,which will be instituted September 10.
Deepfakes,which typically use artificial intelligence to realistically superimpose a person's face onto a pornographic video,started making headlines last year—most notably with the spread of fake porn ofWonder Womanstar Gal Gadot.Creepshots,on the other hand,have been aroundfor a long-ass time.(Ifirst wroteabout a subset of the phenomenon ten years ago.I am old.) In fact,five whole years ago,Jezebel highlightedTumblr's creepshot problem—but now,half a decade later,Tumblr is taking it seriously.Bra-vo.
Tumblr's announcement explains:
Posting sexually explicit photos of people without their consent was never allowed on Tumblr,but with the invention of deepfakes and the proliferation of non-consensual creepshots,we are updating our Community Guidelines to more clearly address new technologies that can be used to humiliate and threaten other people.
The new guidelines—which are meant to eliminate "any ambiguity in our zero-tolerance policy on non-consensual sexual images," according to Tumblr—will include the following instructive: "Don't engage in the unwanted sexualization or sexual harassment of others."
The new guidelines will also "specifically ban the glorification of violent acts or the perpetrators of those acts." The announcement explains,"Not all violence is motivated by racial or ethnic hatred,but the glorification of mass murders like Columbine,Sandy Hook,and Parkland could inspire copycat violence." Along with that,the guidelines will remove previous language that instructed users to only report "negative speech" that rises "to the level of violence or threats of violence." The Tumblr announcement explains,"We believe in a free and open internet but we can't ignore that the internet is being exploited by hate groups to organize,recruit,and radicalize with horrifying efficiency."
Of course,as is always the case with attempts to censor harmful speech online,questions remain.Where will they draw the line between "negative speech" and "hate speech"?How are they going to determine what is,and is not,a deepfake,a creepshot,or any other kind of nonconsensual sexual image?The answer appears to be: Hiring a heck ton of moderators.The post explains,"An overwhelming amount of care and nuance is needed to evaluate reports fairly and accurately,so we've increased the size of our team to review the reports we receive."