‘Take It Down:’ a tool for teens to remove explicit images

    87

    “Once you send that photo, you can’t take it back,” goes the warning to teenagers, often ignoring the reality that many teens send explicit images of themselves under duress, or without understanding the consequences.

    A new online tool aims to give some control back to teens, or people who were once teens, and take down explicit images and videos of themselves from the internet.

    Called Take It Down, the tool is operated by the National Center for Missing and Exploited Children, and funded in part by Meta Platforms, the owner of Facebook and Instagram.

    The site lets anyone anonymously — and without uploading any actual images — create what is essentially a digital fingerprint of the image. This fingerprint (a unique set of numbers called a “hash”) then goes into a database and the tech companies that have agreed to participate in the project remove the images from their services.

    Now, the caveats. The participating platforms are, as of Monday, Meta’s Facebook and Instagram, Yubo, OnlyFans and Pornhub, owned by Mindgeek. If the image is on another site, or if it is sent in an encrypted platform such as WhatsApp, it will not be taken down.

    In addition, if someone alters the original image — for instance, cropping it, adding an emoji or turning it into a meme — it becomes a new image and thus need a new hash. Images that are visually similar — such as the same photo with and without an Instagram filter, will have similar hashes, differing in just one character.

    “Take It Down is made specifically for people who have an image that they have reason to believe is already out on the Web somewhere, or that it could be,” said Gavin Portnoy, a spokesman for the NCMEC. “You’re a teen and you’re dating someone and you share the image. Or somebody extorted you and they said, ‘if you don’t give me an image, or another image of you, I’m going to do X, Y, Z.’”

    Portnoy said teens may feel more comfortable going to a site than to involve law enforcement, which wouldn’t be anonymous, for one.

    “To a teen who doesn’t want that level of involvement, they just want to know that it’s taken down, this is a big deal for them,” he said. NCMEC is seeing an increase in reports of online exploitation of children. The nonprofit’s CyberTipline received 29.3 million reports in 2021, up 35% from 2020.

    Meta, back when it was still Facebook, attempted to create a similar tool, although for adults, back in 2017. It didn’t go over well because the site asked people to, basically, send their (encrypted) nudes to Facebook — not the most trusted company even in 2017. The company tested out the service in Australia for a brief period, but didn’t expand it to other countries. In 2021, it helped launch tool for adults called StopNCII — or nonconsensual intimate images, aka “revenge porn.” That site is run by a U.K. nonprofit, the UK Revenge Porn Helpline, but anyone around the globe can use it.

    But in that time, online sexual extortion and exploitation has only gotten worse, for children and teens as well as for adults. Many tech companies already use this hash system to share, take down and report to law enforcement images of child sexual abuse. Portnoy said the goal is to have more companies sign up.

    “We never had anyone say no,” he said.

    Twitter and TikTok so far have not committed to the project. Neither company immediately respond to a message for comment Sunday.

    Antigone Davis, Meta’s global head of safety, said Take It Down is one of many tools the company uses to address child abuse and exploitation on its platforms.

    “In addition to supporting the development of this tool and having, reporting and blocking systems on our on our platform, we also do a number of different things to try to prevent these kinds of situations from happening in the first place. So, for example, we don’t allow unconnected adults to message minors,” she said.

    The site works with real as well as artificial intelligence-generated images and “deepfakes,” Davis said. Deepfakes are created to look like real, actual people saying or doing things they didn’t actually do.