How Sidechat Accelerated Protests on University Campuses

The anonymous social media platform Sidechat has become a breeding ground for hate speech in the midst of conflicts over free speech on college campuses and the ongoing Israel-Hamas conflict.

In the months following Hamas’ October 7 attack on Israel, conversation on college campuses has been defined by a palpable tension. Increased antisemitic and anti-Muslim rhetoric embroiled numerous universities in free-speech debates. In late April, as the Israel-Hamas War moved into its fifth month, students at Columbia University and other institutions across the US began protesting, calling for a ceasefire. Amid all of this, one platform has served as a locus: Sidechat, a social media app that’s become both a place for dialog about the protests and a breeding ground for hate speech.

Students have been using the app to share memes and voice their displeasure with their administrations’ responses as demonstrations have broken out at Columbia, NYU, Yale, Princeton, the University of Texas, and other places over the past several weeks.

Colin Roedl, editorial page editor of the student-run Columbia Daily Spectator, told Slate on April 22, after a weekend of arrests at Columbia, that students were seeing “calls for solidarity” on the app. The next day, the deans, the board of trustees, and about 3,000 members of the Columbia community signed a statement endorsing “campus safety and academic freedom” and addressing Minouche Shafik, the university president. It contained a link to a folder containing screenshots from Sidechat that showed users asking questions about Zionism and how to join campus encampments.

Hundreds of demonstrators were taken into custody by the New York Police Department on Tuesday at City College and Columbia Universities.

Administrators at other universities, including Harvard and Brown, had tried to tighten Sidechat moderation before the protests because of a rise in accusations of harassment and hate speech from students utilizing the app. “Dehumanizing, racist, homogenizing, (and) hateful” is how one Brown student who is Palestinian, Aboud Ashhab, describes the language on the app. University student Andrew Rovinsky, who is Jewish, refers to it as “a cesspool.”

Toxic remarks and derogatory language are readily used on the app because its distinguishing characteristic is student conversation conducted anonymously (users don’t post with their real names). Because Sidechat is anonymous, Rovinsky claims that what you see there are actual people participating in some of the most disgusting rhetoric you have ever seen.

Sidechat was first introduced in 2022 as a way for college students to discreetly discuss events on school, and it swiftly expanded throughout US campuses. Similar to the initial iteration of Facebook, the application necessitates the use of a university email address for login purposes. Although Sidechat was initially utilized as a gathering place for gossip and group grievances, university authorities have noticed an increase in contentious conversations on the platform and have urged Sidechat to enhance its content moderation.

Although Sidechat and its predecessor Yik Yak have faced criticism for creating an online space that encourages hate speech, the app’s user guidelines specify that content that “perpetuates the oppression of marginalized communities by promoting discrimination against (or hatred toward) certain groups of people” is not permitted on the platform.

In fact, Yik Yak took a four-year break prior to Sidechat acquiring it in 2023 due to a barrage of complaints about racism, bigotry, and violent threats that were making the rounds on the app. In the months after the October 7 incident, hateful remarks have suggested that Sidechat is not all that dissimilar from its predecessor.

Harvard was the first academic institution to request that Sidechat tighten its moderation at the beginning of 2024. Sebastian Gil, one of Sidechat’s cofounders, responded to Inside Higher Ed via email, stating that the app modifies content “more than most (if not all) social media apps.” He cited the app’s 30 member team and its use of machine learning models to “detect bigotry.” (Gil made the identical statement in response to WIRED’s request for comment.)

Following talks with Sidechat leadership “to determine what steps are being taken by Sidechat” to control content on the platform, the Brown administration issued a campus-wide message in March that included instructions for students to report hate speech on the app.

Publicly accessible Brown campus Community Council meeting on March 20 saw campus president Christina Paxson respond to community members’ concerns. “To address the query of whether a university could essentially do away with Sidechat,” she remarked. “The answer is not really, putting aside concerns about censorship, which we take very seriously.” Although other universities, such as the University of North Carolina, have attempted to prohibit the app, their success is limited to the time students spend using their own devices and internet networks.

On April 24, students at Brown University set up a pro-Palestinian encampment on the Rhode Island campus to protest Israel’s actions in Gaza. On Tuesday, they reached an agreement that requires administrators to meet, and vote on, divesting from companies with ties to Israeli interests. On Sidechat, users posted memes with Paxson’s face edited onto Donald Trump’s body on the cover of his book Trump: The Art of the Deal. Students also posted screenshots of comments from the Instagram page of the pro-Palestinian Brown Divest Coalition that called student protesters terrorists and the school “jihadist Brown University.”

According to Rovinsky, as Brown’s campus “grew more tense,” he has witnessed an increase in racist, antisemitic, and anti-Palestinian content on the app within the last six months. He distinguished between two categories of language that he frequently sees used on Sidechat: overt hate speech, such as messages that declare that “all Palestinians are terrorists,” and more nuanced, covert remarks that are less overt, “like employing classic antisemitic tropes and dog whistles.”

These are the items that moderators are missing, according to Rovinsky. Say something like, if you’re not merely calling Jews or Palestinians names. Many people speak in code these days.

Although Ashhab has seen subtle instances of anti-Arab sentiment on the platform, he pointed out that most of the racist material he has seen is straightforward: “They have made statements like ‘all Palestinians and Arabs blow up buses and stab people,’ and ‘all people in Gaza are backward, they all kill women, they’re all rapists.'”

Although the app’s guidelines declare that it does not “allow discrimination against marginalized identities, included but not limited to racism … [and] religious discrimination,” Ashhab wondered if the rhetoric he has encountered is on the platform’s radar at all.

Elizabeth Lokoyi, a junior at Brown University who moderated Sidechat on a part-time basis during her first year’s spring, characterized her employer’s policies regarding the application of the ban on hate speech as “vague.” A large portion of such decision-making is returned to the students using the app.

“The moderator has a great deal of discretion over what constitutes hate speech,” she explains. It was somewhat difficult to identify what set hate speech apart from dubious conversation. However, anything that involves violence, such as threats or disparaging remarks about a particular group of individuals, would be labeled as hate speech and would instead be reported for removal.

There are several restrictions on what schools and universities may do to stop antisemitic and anti-Palestinian discourse on Sidechat because it is an external forum. While some students think that universities have no business monitoring Sidechat, the majority concur that Sidechat can improve its moderation by reducing the amount of AI algorithms used to assess posted content and entirely relying on human moderators.

“I don’t believe Sidechat should be governed by the university. That actually is a task for the organization, according to Rovinsky. “Students shouldn’t be moderators because, quite obviously, we all have personal prejudices about what constitutes hate speech and because we are active participants in that online community.”

In a similar vein, Ashhab thinks that because AI systems are not human, they cannot be adequately educated to weed out subtle hate speech or racist discourse. He claims that “AI can’t just detect all dehumanizing rhetoric and language.” “That’s something that cannot be determined by simply accepting the message at face value or by sifting through keywords.” Moreover, AI has never visited a university campus.