Here’s a number that should stop you for a second.
83 million.
That’s the estimated number of American adults who experienced harassment in online multiplayer games in a single recent six-month period. Not lifetime. Not “at some point.” Six months.
To put that in context: 83 million is larger than the entire population of Germany. It’s more people than voted in the 2000 U.S. presidential election. It’s not a niche problem affecting a small corner of gaming culture. It’s something happening to the majority of people who play online multiplayer games, repeatedly, at a scale most industries would treat as a crisis.
Gaming mostly treats it as a background condition.
I’ve spent months pulling together every piece of published research on gaming toxicity I could find, academic studies, industry reports, think tank analyses, developer transparency data. What follows is the most complete picture I can put together of what toxic gaming actually costs. Not in vibes. In numbers.
How Widespread Is It, Really?
Let’s start with the prevalence data, because the scale is what makes everything else make sense.
The Anti-Defamation League has been running an annual survey on gaming harassment since 2019. Their 2023 report surveyed roughly 2,000 American adults who play online multiplayer games. The finding: 76% had experienced some form of harassment in the previous six months. Applied to the estimated 110 million Americans who play online multiplayer, that’s about 83 million people.
That 76% isn’t an outlier. It’s consistent with other research:
A Unity/Harris Poll survey of 2,929 players found 74% had encountered toxic behavior, up from 68% when they ran the same survey in 2021. Toxicity wasn’t improving on its own, it was getting worse before the current wave of AI moderation tools started deploying.
A 2024 academic study by Kowert and colleagues (n=432) found 82.3% of players had experienced toxicity firsthand and 88.1% had witnessed it happen to someone else. When you’re in a game with someone being targeted, you’re part of that experience too, as a bystander if nothing else.
The thing that makes these numbers feel less abstract: we’re not talking about a player base that’s mostly teenagers who shrug it off. The average gamer is now 36 years old. The people absorbing this harassment are adults with jobs, families, limited free time, and a choice about where to spend it.
Who Is Getting Hit the Hardest
The 76% average conceals significant variation. Some groups experience dramatically worse harassment rates than others, and the gap has been widening.
Black gamers. In 2020, the ADL found that 31% of Black adult gamers had experienced race-based harassment. By 2023, that number had risen to 50%. That’s not a rounding error. That’s a 60% increase in identity-based targeting in three years, during a period when multiple major studios were deploying new moderation infrastructure. The AI tools are helping overall, but they haven’t closed this gap.
Women. The ADL found 48% of women experienced gender-based harassment. A separate 2026 academic study focused on competitive gaming found that approximately 87% of women and transgender and gender-diverse players reported gender-related harassment in competitive environments. The gap between overall averages and competitive mode experiences is significant, competitive ranking systems create conditions that intensify targeting.
LGBTQ+ gamers. GLAAD’s 2024 Gaming Report surveyed 1,452 players. 52% of LGBTQ+ gamers reported experiencing harassment, compared to 38% of non-LGBTQ+ players. More striking: 42% said they actively avoided certain games because they anticipated being harassed there. 61% reported being uncomfortable using voice chat at all. These are people who’ve restructured how they play games, what modes they use, what games they buy, because the environment isn’t safe for them.
Jewish gamers. The ADL data shows 34% of Jewish gamers experienced identity-based harassment in the 2023 survey, up from 22% in 2021. That increase tracks with broader trends in online antisemitic activity during the same period.
Teens. A 2024 Pew Research survey of 1,423 teenagers found 43% of teen gamers had experienced harassment or bullying in online games. The ADL’s youth data is starker: 75% of players aged 10-17 experienced some form of harassment, up from 67% in 2020. And only 39% of youth reports to game companies resulted in any meaningful action.
The pattern across all of these: identity-based harassment is worsening faster than overall harassment rates. The AI moderation tools are catching explicit content more reliably than they’re catching the subtler targeted patterns that make specific groups feel unwelcome. That’s a real gap and an active area of development.
What People Actually Do When It Gets Bad
Numbers about harassment rates are one thing. What people change about their behavior is where the real cost shows up.
The Take This/Nielsen study from 2023 surveyed 2,328 players on behavioral responses to toxic gaming. The findings are worth sitting with:
61% of players said they had decided not to spend money in a game because of how other players treated them. Not because the game was bad. Not because they couldn’t afford it. Because the community was hostile enough that they pulled back their wallet.
60% said they had quit a game session or left a game permanently because of toxic behavior they encountered. Not paused. Quit. Game closed, session over, sometimes account deleted.
72% said they had avoided certain games entirely because of a reputation for toxic communities. They didn’t even try. The reputation preceded the experience.
These aren’t small behavioral shifts. This is a significant portion of the player population making active decisions to spend less time and money in gaming because of how other humans behave in those spaces. That cost doesn’t show up on anyone’s quarterly report, but it’s real.
Riot Games has its own data on this from inside League of Legends. First-time players who experience toxicity in their early sessions are 320% more likely to quit the game immediately compared to first-time players who don’t. Not 20% more likely. Not 50%. Three hundred and twenty percent. One bad experience in the introduction period sends players to a different game permanently.
Research from UC Irvine adds another dimension: players spend 54% more money on games they perceive as having non-toxic communities compared to games they perceive as having toxic ones. Same game, different community reputation, meaningfully different spending behavior. The players who stick around in perceived-safe environments aren’t just playing more, they’re buying more.
GGWP’s analysis across their client games estimates that toxicity drives up to 20% additional player attrition beyond normal churn. One in five players who leave do so partly because of community hostility rather than because they got bored or found something else.
The Mental Health Layer
The research doesn’t stop at behavior. There’s a documented psychological impact that sits under the quit rates and spending changes.
Take This, a nonprofit focused on mental health in gaming, has produced research in partnership with Nielsen on the psychological effects of toxic gaming environments. The findings map onto what you’d expect from any chronic stressful environment: elevated anxiety, increased cortisol, difficulty separating gaming stress from work stress, hypervigilance in online spaces that was supposed to be leisure.
For the players taking the hardest hits, the ones facing identity-based harassment repeatedly, the effect is more severe. The GLAAD data on LGBTQ+ gamers shows a community that has adapted its behavior around anticipated threat. Avoiding voice chat. Avoiding certain games. Playing in ways designed to minimize exposure rather than maximize enjoyment. That’s not a gaming problem. That’s a stress management response to a persistently hostile environment.
Kowert’s 2024 research noted something that stuck with me: 88.1% of players had witnessed toxicity happen to someone else. Bystander effects in online gaming are similar to bystander effects in other contexts, repeated exposure to someone else being targeted creates its own anxiety and disengagement. You don’t have to be the target to be affected.
For adult gamers 28-42, the demographic this site is for, the mental health dimension has a specific texture. These are people using gaming as stress relief from demanding professional and family lives. When the supposed stress relief becomes another source of stress, the calculus changes. The 72% who avoided games because of toxic reputation weren’t making an irrational choice. They were protecting their limited leisure time from environments that had demonstrated they’d make things worse, not better.
The Business Case Nobody Wants to Say Out Loud
Studios talk about toxicity as a community problem and a values issue. Both are true. But the data makes clear it’s also a revenue problem of significant scale, and that’s worth saying plainly, because it’s the argument that moves the most infrastructure dollars.
Let’s put some rough math on it.
If League of Legends new players are 320% more likely to quit after a toxic experience, and Riot acquires hundreds of thousands of new players per year, the toxicity churn rate represents a substantial portion of their long-term player lifetime value loss. Riot has never published a dollar figure on this, but it’s why they’ve invested more in behavioral systems than almost any studio in the industry. The ROI calculation isn’t complicated.
At a platform scale: if 61% of players pull back spending after toxic experiences, and the gaming industry generates roughly $200 billion in global revenue annually, the drag on that number from toxicity-related spending reduction is in the billions. It’s not money going somewhere else in gaming, it’s money that doesn’t get spent on DLC, cosmetics, and subscriptions because the player disengaged.
The UC Irvine finding that players spend 54% more in communities they perceive as non-toxic is the most direct business case in the dataset. It means that the studios investing in real moderation infrastructure aren’t just doing the right thing, they’re building an asset. A community with a reputation for safety retains players longer, generates more spending per retained player, and spends less on user acquisition because word-of-mouth drives organic growth in a way that toxic games simply don’t get.
The publishers who’ve read this data are investing. Riot builds internal behavioral systems more sophisticated than most companies in the space. Activision deploys ToxMod at scale and publishes its results. Microsoft acquired Two Hat and built Community Sift into Xbox’s core infrastructure.
The publishers who haven’t read the data, or who’ve read it and concluded the short-term moderation cost isn’t worth the long-term retention benefit, are leaving money on the table while also creating worse environments for their players. Those two things are connected.
What Progress Actually Looks Like
I want to be fair to an industry that has genuinely moved on this over the past three years.
The before-and-after data from studios that have deployed real moderation infrastructure is real. Call of Duty’s 43% reduction in toxic voice chat exposure. Riot’s 20x improvement in automated text enforcement. Ubisoft’s 50% reduction in flagged messages in Rainbow Six Siege after deploying automated text moderation. GGWP’s work with Predecessor showing 30x more incident detection than manual moderation.
These aren’t press release numbers. They’re published in transparency reports and case studies with methodology notes. The tools work. The question is whether the industry deploys them broadly enough to move the aggregate numbers, the 76% harassment rate, the 83 million affected Americans, rather than just improving individual game environments.
The Unity/Vivox and GGWP partnership is meaningful here precisely because it extends real moderation infrastructure to games that couldn’t previously access it. The ADL report showing toxicity getting worse from 2020 to 2023 predates the current wave of AI moderation deployments. If the same survey gets run in 2026 and 2027, those numbers should start moving.
They haven’t moved yet. The 2024 and 2025 data don’t show the broad population-level improvement you’d expect if the tools had reached enough of the ecosystem. But the tools are getting there.
The TAG Toxicity Funnel
I’ve been thinking about how to visualize the data in this article in a way that makes the business and human case simultaneously. Here’s the model:
The Toxicity Funnel
100 online multiplayer players enter a game environment. Across a six-month period:
- 76 experience some form of harassment (ADL 2023)
- 60 modify their behavior as a result, muting, avoiding modes, changing how they play
- 52 stop spending as freely in games where harassment occurs (subset of Take This/Nielsen 61%)
- 43 avoid certain games entirely based on community reputation
- 28 quit a game permanently at least once due to toxicity
- ~15 who belong to a targeted group (women, LGBTQ+, racial minorities) experience identity-based targeting
At each stage, you’re losing player engagement, spending, and long-term retention. The players who make it through without a significant negative experience, roughly 24 out of 100, are the ones who stay longest and spend the most. Studios that invest in moderation infrastructure are fighting to widen that 24 to a larger number.
Every percentage point improvement in the harassment rate doesn’t just make things better for players, it moves players from the left side of this funnel toward the right. And on the right side is where the spending, the retention, and the word-of-mouth growth live.
What This Means
Toxicity isn’t a side problem in gaming. It’s a structural feature of how most multiplayer environments have been built, affecting the majority of people who use them, with measurable costs at the individual, community, and business levels.
The players most affected aren’t abstractions in a survey. They’re the adults 28-42 who game for stress relief and find the stress follows them in. They’re the women who’ve stopped using voice chat because the math doesn’t work out in their favor. They’re the LGBTQ+ players avoiding entire games because the community has made clear the welcome isn’t genuine. They’re the Black and Jewish players watching identity-based harassment numbers climb even as overall tools improve.
And they’re the new players, the ones who tried a game, ran into something ugly in the first few sessions, and decided gaming wasn’t for them. Three hundred and twenty percent more likely to leave permanently, gone before anyone had a chance to show them what gaming could actually be.
The data tells a story about what’s being lost. The tools exist to start telling a different one.
Have data we should add to this? Academic research, industry reports, or your own experience that the numbers here are missing? Tag me in the TAG Discord or drop it in the comments. This report will be updated as new data becomes available, and we’re tracking the ADL survey release closely to see whether the AI moderation wave is moving the numbers.
