Bluesky’s moderation woes continue as users threaten to leave the site in protest of its failure to flag slurs in account usernames.
Many users — particularly Black users — are frustrated that Bluesky hasn’t apologized for allowing racial slurs to slip through its moderation tools even though they violate the platform’s community guidelines. It’s the latest miss for the company, which has been under backlash for being slow to crack down on hate speech and threats against marginalized communities.
The platform is also under fire for removing numerous racist, ableist and transphobic slurs from its list of flagged words in a contentious update last week.
“Our community guidelines published yesterday reflect our values for a healthy community, and we’re working on becoming better stewards every day,” Bluesky CEO Jay Graber said in a post on Saturday.
Last week, users reported an account that used a racial slur as its username. The account had been active for 16 days before users flagged it. Bluesky removed the account the same day.
“User handles that are slur words are a form of harassment,” Bluesky said in a post. “We’ve deployed a change so that these handles can no longer be created in the app.”
Bluesky updated its banned word list, which includes slurs, expletives and celebrity names that cannot be used as usernames when creating a new account. But the change didn’t account for already existing accounts, and one user was able to change their handle to a racial slur hours after the update.
Users flagged that account, and questioned why it was able to slip through Bluesky’s banned words filter.
As many users pointed out on GitHub, the update also removed numerous racist, ableist and transphobic slurs from the list of words that users are not allowed to use in their handles. The update prompted fierce debate in the GitHub comments, in which some users argued that certain words that are considered slurs in English have innocuous meanings in French and Spanish. Others noted that the list only prevents users from including flagged words in their usernames, not from using them in posts. Bluesky eventually locked the thread and marked it as “too heated.”
“The next time I hear someone say code can’t be racist or ableist, I’ll just point them to this commit,” GitHub user siobhandougall commented on the update before Bluesky locked the thread. “Anything to say about it? Or are you just gonna lock comments so you can all pretend there are no consequences?”
In protest of Bluesky’s missteps and lack of apology for failing to implement a slur filter, swaths of users threatened to leave the platform. Rudy Fraser, who created Blacksky — a custom feed for Black Bluesky users — said he would delete his account if Bluesky didn’t respond.
“Someone else can host the feed if they’d like,” he posted. “Bluesky’s silence has made y’all bold.”
Others vowed to stop engaging with the platform entirely. The Twitter-famous Dril announced a “posting strike from here until they make every one not racist or whatever.”
Some users did follow through. Aveta, a software engineer who shaped Bluesky’s usership by inviting hundreds of Black Twitter users during Bluesky’s early days, mourned the shrinking Black community on the app.
“All the beautiful people that [I] helped invite over that left,” she posted. “Literally this is my community, black tech. damn man.”
The next day, Bluesky announced updates to its terms of service and community guidelines. The community guidelines forbid users from using the platform to “break the law or cause harm to others,” the company said in a post. Users are also expected to “treat others with respect,” and Bluesky does not allow conduct that “targets people based on their race, gender, religion, ethnicity, nationality, disability, or sexual orientation.”
In the comments, users pressed the company on whether it’ll enforce the community guidelines, and questioned whether Bluesky will hire more human staff to its trust and safety team, instead of relying on automated moderation.
The company did not reply to user comments, but posted a thread the next day, stating that racism and harassment have “no place on Bluesky.” The company also said it invested in expanding its trust and safety team, improving and clarifying policies and prioritizing moderation tools.
“On Wednesday, users reported an account that had a slur as its handle. This handle was in violation of our community guidelines, and it was our mistake that allowed it to be created,” Bluesky stated in the thread, which was posted after midnight EST. “40 minutes after it was reported, the account was taken down, and the code that allowed this to occur was patched. To make this a great place as we grow, we’ll continue to invest in moderation, feedback, and support systems that scale with the number of users on the app.”
Bluesky did not issue a public apology. Embittered users don’t expect it from the platform.
“I’m not sure why anyone is waiting on the Bluesky staff to apologize,” software developer Angie Jones posted. “Obviously, they aren’t sorry, nor regretful. You think they forgot to exclude that word?! Of course not.”
At least one Bluesky developer has apologized. Bryan Newbold, who was widely tagged on GitHub for publishing the banned words update, explained that the update replaced the “publicly contributed list of slurs” with an “emergency list” that was based on identifying slurs targeting Black users. The emergency list also included antisemitic terms, he said in a GitHub comment. The banned words list was a “temporary measure” that has since been replaced with a “more complex mitigation” that will “surely need to be revised over time.”
Newbold also addressed the criticism in a Bluesky thread on Sunday.
“I have made decisions, and made mistakes. those have caused harm to real people, including Black folks, including really great [and] knowledgeable Black folks who supported Bluesky,” he wrote. “I’m sorry. feel pretty bad about it. it sucks. re/earning their trust, and everybody else’s trust, will be hard.”
Originally published at techcrunch.com