The enormous work messaging platform Slack quickly reversed course yesterday, promising to revise a brand-new direct message feature that could have been misused for harassment.
Added to the company’s “Slack Connect” product—which lets enterprise users share messages with contract workers and third-party partners outside their company—the new “direct message” feature allowed paying Slack users to message anyone outside of their company or organization, so long as they had another person’s email address. The messages came attached to an invite, but as many tech news outlets and concerned online users noted, there was no way for recipients to block the invites, or to block the content of the messages that came attached to the invites.
As Twitter product employee Menotti Minutillo said on Twitter, the implementation of Slack Connect DMs meant that malicious users could send repeated DM invites with harassing language, and that Slack would also email the DM’s recipient with the invite, including the harassing language. DM recipients would also have trouble blocking those emails as they came from a generic email address, too, Minutillo said.
Further, according to TechCrunch, the Slack Connect DM feature is opt-in at the organizational level, meaning that individual employees could not, alone, overwrite their company’s decision, should it choose to enable the feature.
Less than 24 hours after Slack Connect DM’s full release, Slack realigned. According to Slack Vice President of Communications and Policy Jonathan Prince, the company will disable the capability to customize messages that are attached to Slack Connect DM invites.
Prince’s full statement is as follows:
“After rolling out Slack Connect DMs this morning, we received valuable feedback from our users about how email invitations to use the feature could potentially be used to send abusive or harassing messages. We are taking immediate steps to prevent this kind of abuse, beginning today with the removal of the ability to customize a message when a user invites someone to Slack Connect DMs. Slack Connect’s security features and robust administrative controls are a core part of its value both for individual users and their organizations. We made a mistake in this initial roll-out that is inconsistent with our goals for the product and the typical experience of Slack Connect usage. As always, we are grateful to everyone who spoke up, and we are committed to fixing this issue.”
Slack’s quick work to fix the problem is appreciated, but it is curious that the company did not catch the problem before the full rollout. The company has already faced complaints about the limited features in the free version of its platform, which allows users to visibly show harassing language without even having to actually write and send messages. This is because Slack automatically sends notifications when new users join a thread, so if those new users stylize their username to be an insult, then the users in that thread will receive a notification that includes that language.
Further, the problem of harassment on messaging platforms is far from new. On the Lock and Code podcast, when we spoke with Electronic Frontier Foundation’s Director of Cybersecurity Eva Galperin, Galperin warned about this very issue.
“Primarily, the onus for making safe platforms, is on the makers of the platforms,” Galperin said. “And so, if there are people who are listening to this podcast, who are developing software or who are developing platforms or services for commercial use, I encourage them to think about how their tool will be used for harassment.”
Galperin provided specific guidance for any platform with messaging capabilities. She said that those platforms should make it possible for users to not use their real names, and for users to block other users or to mute certain keywords. This setup, Galperin said, is beneficial for both the user and the company.
“If you give the power to the users, then they can decide what is harassment and what is abuse, and it really takes the onus off the platform to be judge, jury, and executioner for every communication that somebody has online.”
Unfortunately, Slack users could not block users—and in fact the company has pushed back against such a feature for years—or mute keywords, and users would have trouble filtering out emails from Slack’s generic email addresses that included the DM invites and the accompanying messages.
These may sound like high-level discussions that are difficult to forecast, but there is actually a far simpler way to look at the problem. To borrow the words of Twitter user @geekgalgroks, a developer and accessibility advocate:
“Seriously with every new messaging system and feature ask yourself if people can send unsolicited dick pics and if those receiving them can block the sender.
Because it will happen.”