Jason Dorfman/MIT CSAIL
Amy Zhang
- MIT researchers released a new tool called Squadbox that aims to combat online harassment.
- The tool allows users to let a group of friends, or "squad," filter out harassing emails.
- Squadbox is only available for email right now, but it could soon expand to other private messaging systems, such as Direct Messages on Twitter or Facebook Messenger.
Researchers at the Massachusetts Institute of Technology on Thursday unveiled an unconventional approach to moderating online harassment: let your friends do it.
"Squadbox," a tool created by a team at MIT's Computer Science and Artificial Intelligence Laboratory, allows users to assemble a "squad" of friends to monitor and filter out harassing emails. Anyone can sign up at squadbox.org.
"There's a lot of discussion about platforms and algorithms managing harassment or paying an army of content moderators to look through it for us. And I think those things are valuable," Amy Zhang, a PhD student and lead author on the project, told Business Insider. "But I also think there's room for a tool like Squadbox to live alongside these things. I think it's important that people have more control over their personal experience instead of every single person on a platform following the same set of content guidelines."
With Squadbox, users set up filters indicating which emails they want forwarded to their "squad." Once an email arrives, a moderator logs into Squadbox and decides which emails are harassment, and which can be forwarded back to the person's inbox. The program has whitelists and blacklists, so users can designate if messages from certain senders will bypass their squad and go straight into their inbox or if an email will automatically be rejected.
Users can also give members of their squad permission to put emails in a specific folder if they want to look at them later or to redact certain parts of emails.
Amy Zhang Squadbox's interface
"We found that everyone deals with harassment differently and people have very different ideas of what is harassment, so we wanted to make a tool that was as flexible as possible," Zhang said.
Beyond individual use, Zhang sees Squadbox being used to protect those whose jobs require a public presence online. In academia and newsrooms, for example, "squads" could consist of editors, department chairs, or other managers that would protect academics and journalists from harassment.
Squadbox is only available for email right now, but the team at MIT says they're working on expanding it to other private messaging systems, such as Direct Messages on Twitter or Facebook Messenger.
Zhang and her team found that since email isn't public, harassers might feel less accountable. Additionally, while users can block certain email addresses, there is nothing stopping harassers from creating a new email address and starting the harassment again.
"One of the reasons we chose email is because there's actually very little support for people that are getting harassed by email," Zhang said. "On Twitter, people talk about how fruitless is might be to report accounts to Twitter, but with email, there's no such reporting."
Zhang, whose research involves creating tools to improve online discussion, and her team interviewed 18 people who have been harassed online before building Squadbox. They ranged from people who were being harassed by an ex partner to famous YouTubers who encountered coordinated harassment campaigns.
She first got the idea for Squadbox after a few interviewees mentioned they already give their passwords over to friends so they can moderate their emails. Zhang thought this was intriguing, but she wanted to make a more secure way of doing this - or "friendsourcing."
When Zhang tested Squadbox on five pairs of friends, she found it helped with fears of privacy and allowed for more tailored decisions. But they also found slow response times posed a serious issue. Having multiple people in your squad could potentially help, Zhang said.
In interviews with friends of people who were being harassed, Zhang noticed most of them expressed a desire to help, but didn't know how. At first, Zhang was concerned that reading harassing emails, even if it's not directed at you, would still be harmful. She didn't want to further spread harassment to more people. But her research found that reading someone else's harassment isn't as bad as reading your own.
"We found that if you weren't being targeted, people could read through it and clean it up for their friends," she said.
Plus, Zhang pointed to small design features she hopes will mitigate the effects of reading other people's disturbing emails, such as alerts reminding squad-members to take a break if they need it or ways the person getting harassed can thank their squad inside the program.