Sick of Dick Pics? A New AI Filter Claims It Identifies and Blocks 99 Percent of Unwanted Penises On Twitter

According to a 2017 YouGov survey, 78 percent of millennial women (or three in four ladies) have been sent an unsolicited dick pic in what is refered to as “cyberflashing,” a modern form of sexual harassment. To curb the proliferation of unwanted dicks, three developers created an AI filter for Twitter called “Safe DM,” as BuzzFeed reports. Kelsey Bressler, one of the three developers, told the publication that Safe DM “will block up to 99% of dick pics,” and while users must allow the extension to log into their Twitter account, she swears “the developers can’t read your DMs, but can see if a message contains a dick pic or not.” Hmm.

It seems as though no one should be expect to forfeit individual privacy in order to curb harassment, but I digress. BuzzFeed tested out Safe DM, which launched today (Valentine’s Day, heh) to see if it truly could analyze dick pics and remove them. For the most part, it seemed to work—penises would be removed and the perpetrator would receive a message that said “⚠️That image was NSFW⚠️. Deleted! ❌.” But there were some flops: dicks in condoms weren’t flagged, but dildos, dick-shaped lipsticks, plants, and other phallic objects were removed despite not actually being genitalia.

While I do appreciate that someone out there is working to make the online experience a bit less treacherous and full of harassment, it doesn’t seem fair that users are expected to make themselves more vulnerable by turning over account permissions to do so. Shouldn’t this be on Twitter to flag and delete? What are they up to? And what about creeps who airdrop dick pics to unsuspecting strangers in public? If a culture of cyberflashing is a gaping wound, Safe DM seems like a tiny bandaid. But given the total lack of oversight on social media, I don’t think I’m alone in saying I’ll take what I can get.