�鶹������ý

This article in , a collaboration among , , and .

What goes on the internet stays on the internet. Anything you say publicly—on a blog, in a Facebook post, in a tweet—could be shared far beyond the audience you imagined, and that’s wreaked havoc on countless occasions: Scores of celebrities have come under fire for their . It can destroy careers for noncelebrities too, like Justine Sacco, who was famously fired after her insensitive tweet .

It’s no wonder, then, that people have moved toward smaller, more intimate online communities where conversations aren’t as public. “People are more cautious of having a permanent record of what they’ve shared,” Mark Zuckerberg wrote in a announcing the platform’s shift toward prioritizing closed groups and other private communications. Twitter, too, is introducing ways for users to . These more private spaces often require a special invitation, curated by moderators, and posts made in them can’t be viewed by or shared to people outside the circle. In addition, many groups enforce privacy through norms and etiquette: Like the rules of Fight Club, members shouldn’t talk about the group with nonmembers, and most definitely should not share things from this group with anyone outside of it. Text messages abide by similar rules. I cherish the implicit trust that should our conversation about my “which March sister are you” quiz results turn toward more private topics, those words won’t leave my screen or my friends’. But one thing challenges all of that: screenshots.

Sharing screenshots has become a normal part of our digital behavior.

Screenshots reflect the itch to document a moment that evoked emotion: amusement, incredulity, anger (or maybe all three). Scrolling through my own phone, I see screenshots of Instagram comments that made me laugh, a threatening email from an irate reader, and a weird post from my neighborhood group about an aggressive crow. Each of those came from a closed or private online space, but I wanted to make sure I had some record of them just in case. Some screenshots have never left my phone and are only for me, while I’ve sent others (usually the funny ones) to friends I think would enjoy them.

Sharing screenshots has become a normal part of our digital behavior. Once-private or semi-private messages make the rounds online, entertaining us and stirring up drama. When someone makes a claim and insists they , they’re often screenshots—a series of Instagram DMs, text messages, private social media posts. There are entire online subcommunities dedicated to private drama made public via screenshots; Facebook groups like “i am but a simple moth to a dumpster fire” and “” or subreddits like are built on screenshots of silly exchanges and ridiculous arguments.

The popularity of shared screenshots is a reflection of our slowly eroding private sphere. Try as we might to maintain privacy in personal communications, what the recipients of our data and messages do with our information is largely out of our hands. Take, for instance, Facebook’s Cambridge Analytica scandal: It revealed that users who allowed the This Is My Digital Life app to their data compromised not only their own data, but . (Apps like did the same, allowing third-party companies to harvest a huge amount of user data.) On a smaller scale, sharing with someone—even someone you trust—can backfire spectacularly; an revealed that Jeff Bezos’ texts and photos with girlfriend Lauren Sanchez were compromised by Sanchez herself, when she shared screenshots with her brother.

And now, screenshots might become fodder for researchers. In January, Stanford scientists announced a new endeavor called the Human Screenome Project, which takes screenshots of participants’ phone use every five seconds. The idea is that the screenshots will provide a more nuanced look at “screen time” and the myriad ways we use our phones. On the project’s , a clip shows screenshots representing three minutes of a user’s activity, which includes Google searches, Spotify playlists, scrolling through Instagram, and texts between friends. The researchers say this data will be . Technically, as long as participants consent to their data being recorded, standard research ethics wouldn’t require researchers to do much else to protect users’ privacy. But required research ethics scratch only the surface of what is morally right. (Another example: It’s not technically a violation of ethics to use data from “public” sources like YouTube videos or tweets, but people who in research are often put off by it.) And it’s not only study participants’ data that will be collected—anyone they text or message with can have their words screenshotted too. If I were a close friend or family member of someone participating in this project, I’d want to know if my private conversations were being archived somewhere and potentially viewable by researchers.

Unfortunately, there’s no script for telling your friends and family that you’ve shared their info with an app, or that you’re allowing researchers to access your data. Until we develop social norms for this type of situation, there’s one feature that could help: screenshot notifications. At the very least, it’s high time to think about the screenshot as a privacy liability and consider limiting its power, through app settings and social acceptability.

This is already a regular feature in some apps. Snapchat gained popularity for the ephemerality of its posts—stories stay up for 24 hours, and photos sent directly to another user disappear as soon as the recipient has seen them. To protect that sense of privacy, the app users if someone has taken a screenshot of their content. Instagram users if someone took a screenshot of a photo they sent directly to another user, but that’s no longer a feature.

If this were standard practice across platforms—on, say, Facebook, WhatsApp, Tinder, and smartphones’ native texting apps—people might think twice about taking screenshots without permission. The motivation isn’t always nefarious. I have a folder on my phone of screenshots of funny or heartwarming conversations I want to save for later. But I have also taken and received my fair share of screenshots where I suspect the original poster would be a little upset to learn their content or conversations didn’t stay private: strangers’ terrible dating profiles, annoying posts in neighborhood Facebook groups, text exchanges a friend wants help deciphering. An alert can’t prevent people from taking screenshots and sharing them, but it could deter them. It could also encourage more dialogue between users about data sharing; a notification that says “So-and-so took a screenshot of your photo” can open up a conversation about what they intend to do with that snap, or allow a user to just block someone who consistently creeps on their posts.

Snap says its users have “come to expect” that screenshot notification. “We’ve gotten feedback from users that they really love that feature,” says Katherine Tassi, Snap’s deputy general counsel for privacy. “It’s really a contract between users, who know if they take a screenshot, the other individual is going to get notified.”

Snap’s screenshot alert shows that the absence of such a feature on other apps isn’t a matter of technical difficulty. It’s not complicated for apps to build such a feature into their platforms, at least on mobile devices. Apple includes in its Software Development Kit for iOS that developers can use to identify when a screenshot is taken, so simple detection is “trivial,” says app developer Kevin Donnelly. The next steps—reporting the screenshot to the app’s server and then sending a notification to the affected user—should be fairly straightforward as well. A screenshot alert might be more difficult to execute on Android apps, says Donnelly, because of how flexible the operating system is, but in lieu of screenshot notifications, developers might be able to block screenshots altogether. Similar code could also block screen recording on phone operating systems, another popular way to capture on-screen content. If screenshots generated notifications, screen recording might become a more popular workaround, so it seems wise to anticipate how to control recording as well.

To make this all work, apps would need to encode data in a way that allows for easy identification of whose content is being screenshotted. On Snapchat, for instance, photos display after you click on a user story or open a message, so the app should easily detect exactly whose content a user is looking at. Scrolling through a feed is more complicated. As you read through Facebook’s newsfeed, you see a hodgepodge of posts from friends, individuals posting in groups, or ads. That makes things “harder, but not very hard,” says Donnelly. “You can fairly easily get an idea of what components are on the screen at the time of the screenshot and look at the underlying structured data to identify which users’ content is being displayed.” As long as developers are structuring their data properly to leave a “paper trail” of what users are looking at when they take a screenshot, notifications are certainly possible.

While notifications or screenshot blocking might be a deterrent, they’re obviously not foolproof. Introducing these features might deter the casual screenshotter, but the truly motivated can always find ways around them. The more tech-savvy could write code to get around anti–screen capture features, but there’s also an easy, low-tech workaround: Just take a photo of a screen with another screen. Snap says it’s aware that people try to get around the app’s screenshot notification feature (even publications like have run guides on how to screenshot in secret) and remains vigilant about those strategies so they can fix any potential loopholes users discover. On Android, in particular, the customizability of the operating system means that motivated individuals can get around app code. “Someone with enough time might be able to configure their phone in a way to get around any system the developer implemented,” says Donnelly. And there may be cases where we ��Dz�’t want to block surreptitious screenshots; it’s a good thing when someone can secretly screenshot evidence to hold a serial harasser or abusive partner accountable, or “receipts” to trace . The real solution lies beyond code: We need the norms and language to deter malicious screenshots. But in the meantime, alerts could help.