By Dr. Oleksandra Gudkova,
University of Vienna
“We live in a time where the dangers children face online have never been greater, and every parent deserves to know their child is protected”
This warning, issued by the Attorney General Paxton (Texas, USA), leading a recent case against Discord, captures a growing concern across the gaming industry: children are entering digital play spaces that promise safety, but whose moderation systems often fail to protect them in practice.
Online games are often presented as safe social spaces for children. Major gaming platforms promote parental controls, AI moderation tools, reporting systems, and communication filters as evidence that harmful behavior is being actively managed. Parents are encouraged to believe that while risks exist, moderation systems are capable of keeping children safe.
But a growing number of investigations, lawsuits and regulatory actions suggest a different reality: moderation systems in online games are consistently failing to protect children from harassment, grooming, exploitation and abuse. The problem is not just that moderation occasionally breaks down. The deeper issue is that moderation systems are being expected to govern communication environments they were never realistically capable of governing at scale. Modern multiplayer games are no longer just games. They are massive communication platforms built around continuous interaction. Text chat, voice communication, private messaging, player networks, livestreaming and user generated content have become central to the gaming experience. These systems are intentionally frictionless because engagement depends on players communicating continuously and easily with strangers. From a business standpoint this model is very successful. But from a moderation standpoint it creates environments which are extraordinarily difficult to govern.
One of the first examples that comes to mind is the case of Roblox, an extremely popular gaming platform for children. Roblox has been criticised multiple times over grooming, predatory behavior, and exposure to inappropriate content through its communication systems. Reuters reported in 2026 that Roblox agreed to pay $23 million to settle child safety investigations brought by U.S. states over allegations that it failed to provide minors enough protection from exploitation and harmful interactions. After years of scrutiny Roblox introduced stricter age-based accounts and restricted communication onto its platform. What makes the Roblox case important is not just that harmful interactions occurred – the harm occurred despite there being filters on chats, reporting mechanisms, and parental controls. The safeguards were there, but they could not prevent children from being reached by harmful individuals. This same pattern has also emerged on Discord, a communication platform widely used across the gaming ecosystem (150 million people worldwide). Since 2023, when NBC News published their investigation “Child predators are using Discord, a popular app among teens, for sextortion and abductions”, the platform has been facing a number of lawsuits for enabling child grooming and exploitation. The investigation revealed that since the platform’s creation in 2015, at least 35 child abduction, grooming, or exploitation prosecutions involved communications via Discord, and 165 child sexual abuse material prosecutions involved the platform. Additionally, NBC News identified hundreds of active Discord servers promoting child exploitation. Discord also showcases another challenge for moderating: much of the interaction between users takes place in semi-private spaces that make it difficult for monitoring to take place consistently.
Researchers have also noted that voice communication creates significant enforcement challenges when trying to document and review harmful behavior (Jiang, 2019). Mainstream multiplayer games like Fortnite and Call of Duty have been dealing with the same issues for years. Voice chat systems in these games have been criticised for allowing younger players to be exposed to harassment, hate speech, sexualised behavior and predatory interactions. Companies now rely heavily on AI to help them moderate their environments. However AI moderation technologies frequently fail to understand context, detect subtle grooming behavior or intervene quick enough to stop escalation of harmful behavior.
The response to these failures by the industry has been generally limited to additionally developing moderation technology rather than question whether these communication environments can even realistically be moderated at all. This is where the broader governance problem exists. Moderation systems are often used as evidence that platforms are acting responsibly. Filters exist. Reporting tools exist. AI moderation systems exist. But the visibility of moderation systems does not equal effective protection. In practical terms moderation usually functions more as a signal of responsibility (safety tool) then as actual protection.
Recent legal action taken against Meta in New Mexico reflects growing skepticism toward industry claims about moderation effectiveness. Reporting by The Washington Post and the Guardian document how a jury found Meta did not sufficiently protect children concluding that its platforms enabled environments where predators could target minors. While Meta is not a gaming company, the case demonstrates a shift: courts and regulators are starting to question if platforms overstate the effectiveness of their safety systems. At its core the problem relates directly to the business models used by digital platforms. Friction introduced through effective moderation restricts communications, slows interaction and limits growth. Therefore moderation systems are designed to introduce minimal disruption rather than maximise protections.
Children are on the receiving end of this contradiction. Younger players tend to believe what strangers tell them, they tend to underestimate the risk and seek validation through online interactions. In a system where open and continuous communication is encouraged, younger players’ vulnerability will be exploited at an unprecedented scale.
The gaming industry has created a perception that digital games are community-based and enjoyable, in part this is true. However, the persistent issues of grooming, harassment and abuse throughout all aspects of gaming communities suggest that there is something structurally wrong with how moderation works currently.
As concerns about child safety continue to grow, the challenge facing the gaming industry extends beyond developing better moderation tools. It requires a broader reconsideration of how digital play spaces are structured, who bears responsibility for protecting young users, and whether systems built for engagement can also be designed with safety at their core.
The question, then, is not simply whether moderation can be improved, but whether moderation alone can ever make these spaces truly safe for children.
This blog post was written as part of the Horizon Europe project GAMEHearts, which investigates the cultural, social, and economic dimensions of the European video game industry. To learn more about the project, visit gamehearts.eu.
References:
Jiang, J. A., Kiene, C., Middler, S., Brubaker, J. R., & Fiesler, C. (2019). Moderation Challenges in Voice-based Online Communities on Discord. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), 1–23. https://doi.org/10.1145/3359157
Leave a Reply