After a new report accused Steam of being a breeding ground for extremism, a US Senator has threatened to hit Valve with government action if they do not immediately work to remove ‘hateful content’ from their online storefront.
Assembled by the Anti-Defamation League (ADL) and titled ‘Steam-Powered Hate: Top Gaming Site Rife with Extremism & Antisemitism‘, the report in question was assembled by the organization’s Center on Extremism, who supposedly analyzed “458+ million profiles, 152+ million profile and group avatar images and 610+ million comments on user profiles and groups” on Steam.
In doing so, the ADL claimed to have found millions of examples of extremist imagery and text in each of these areas – which, though it sounds like a lot, should be noted is only a miniscule percentage of Steam’s overall content.
For example, the ADL states that the platform hosted “1.83 million unique pieces of extremist or hateful content”, which boils down to just 0.15% of all content in the Steam Community.
Further, they found that that 1.5 million unique users – or less than 0.33% of profiles – “used at least one potentially extremist or hateful symbol, copypasta or keyword on the platform”, with the organization defining a copypasta as “blocks of text that are copied and pasted to form images or long-form writing.”
Their report also noted that roughly 1.18 million comments contained “unique instances of potentially extremist and hateful copypastas”, amounting to less than 0.2% of all comments across the platform, and that 827,758 avatars “contained extremist or hateful symbols”, accounting for less than 0.545% of all such images.
In explaining their methodology, the ADL explains that the used their ‘HateVision’ AI to analyze the above data in search of specific hate symbols, keywords, and copypastas (though touted as being able to execute its task with “high precision”, the organization admits in the appendices that the AI not only struggles with complex symbols and caricatures, but also lacks facial recognition capabilities, meaning their findings do not represent the entirety of the hateful content on the platform).
Notably, it seems the ADL’s report treats all instances of hate symbols, regardless of a user’s possible ironic intent, as being an endorsement of the related ideology.
Even so, out of the 996,808 “extremist images” analyzed, the report asserts that the most common one was Pepe the Frog – who they consider a hate symbol, albeit one that depends on context – at 54.6% (544,267 images), with swatikas placing second by a wide margin at 9.1% (90,898).
Though the ADL admits that a normal looking ‘Sad Pepe’ was used as one of the AI’s training images, their report affirms that “Because some of the symbols included in this analysis are also used in non-extremist contexts or are visually similar to non-extremist symbols, we sought to prevent false positives by training only on extremist instances of these symbols.”
Turning their criticisms to Steam’s lack of action in removing the ‘hateful content’, the ADL declared, “While Steam appears to be technically capable of moderating extremist and hateful content on its platform, the spread of extremist content on the platform is due in part to Valve’s highly permissive approach to content policy.”
“In rare notable cases, Steam has selectively removed extremist content, largely based around extremist groups publicized in reporting or in response to governmental pressure,” they added, citing a 2018 Kotaku report. “However, this has been largely ad hoc, with Valve failing to systematically address the issue of extremism and hate on the platform.”
To this end, they also chastised how “Steam’s public-facing content policy includes no mention of hate or extremism” – However, this accusation blatantly ignores the fact that Steam’s Terms of Service forbids content or conduct that is illegal, “racially or ethnically offensive,” “harasses, threatens or embarrasses others, or promotes discrimination, bigotry, racism, hatred, harassment or harm against any individual or group,” or promotes or threatens violence against any person or entity.
Likewise, the platform’s Online Conduct and the End User License Agreement also explicitly forbid abusive and offensive language, as well as harassment.
Unsurprisingly, this is not the only major hole in the ADL’s report.
In its opening paragraph, the report recalls the story of the August 2024 mass stabbing at a cafe in Turkey as an example of an instance wherein a user, in this case the perpetrator, was “posting hateful content on the platform” before ultimately having “engaged in an act of offline violence.”
However, this turns out to be nothing more than a cheap appeal to emotion, as per the ADL’s own admission, “there is no evidence to suggest that the attacker was directly inspired by extremist content on Steam”.
As such, the only relation between the attack and Steam, is that the attacker decided to use the platform as his soap-box-of-choice to espouse his own hateful beliefs.
And in a subsequent appeal to authority, the ADL pointed to how “researchers, media outlets, policymakers and governing bodies” have issued similar warnings about Steam’s content.
But upon closer inspection, it becomes clear that three of the authorities they cited – the New York University Stern Center for Business and Human Rights, the European Union, and US Senator Maggie Hassan – all predicated their arguments on the ADL’s own past research, such as their 2022 report on “Hate and Harassment in Online Games”.
In drawing their report to a close, the ADL provides a list of recommended actions for both Steam and US lawmakers to take in service of cleaning extremist content off of the platform.
For Steam, the organization calls for them to “adopt policies prohibiting extremism”, “adopt policies prohibiting hate”, “enforce policies accurately at scale”, revamp their “content moderation practices to close loopholes”, and “consult regularly with civil society groups, academics and researchers from a broad cross section of positions, including civil rights and civil liberties groups, and especially seek out and use their advice and expertise to shape platform policies that may impact the experience of vulnerable and marginalized groups.”
Meanwhile, the ADL asks lawmakers to “prioritize transparency legislation in digital spaces and include online multiplayer games”, “Enhance access to justice for victims of online abuse” (which they see taking the form iof “legislation that holds perpetrators of severe online abuse accountable for their offenses at both the state and federal levels”), “establish a National Gaming Safety Task Force”, and provide federal funding “that enables independent researchers and civil society to better analyze and disseminate findings on abuse.”
But despite its flaws and borderline-protection-racket “recommendations”, the ADL’s report prompted action from Senator Mark Warner (Virginia, Democrat), who wrote a letter directly to Valve CEO and Founder Gabe Newell criticizing his company for taking little action against Steam’s “hate and extremism” problem, especially given how “in 2022, Valve received a Senate letter identifying nearly identical activity on your platform.”
“As Black Friday and the holiday buying season approaches, the American public should know that not only is Steam an unsafe place for teens and young adults to purchase and play online games, but also that, absent a change in Valve’s approach to user moderation and the type of behavior that it welcomes on its platform, Steam is playing a clear role in allowing harmful ideologies to spread and take root among the next generation,” said Senator Warner. “Valve must bring its content moderation practices in line with industry standards or face more intense scrutiny from the federal government for its complicity in allowing hate groups to congregate and engage in activities that undoubtedly puts Americans at risk.”
In seeking specific action from Valve, the Senator ultimately gave Newell a deadline of December 13th to provide him with direct answers to a number of questions “and not a narrative that attempts to answer multiple”.
Numbering ten in total, these questions range from inquiries into “how many human content moderators work for Steam”, “what commitments will Steam make to ensure that it has meaningfully curbed [its hateful content” by this time next year, and “how frequently does Valve evaluate its content moderation practices related to extremism”.
As of writing, neither Valve nor Newell have released a public response to either the ADL’s report or Senator Warner’s inquiry.