In 1997, a Quake player named Dennis Fong was so dominant that he won John Carmack’s Ferrari in a tournament. He was so recognized in the professional gaming world that when Dota needed a hero name for an aggressive support character, they named it Thresh, Fong’s gaming alias.
Dennis Fong more or less invented professional esports. He was the first gamer to be covered by mainstream media as a professional athlete, the first to get sponsored, the first to win a car at a gaming tournament. If you’ve been gaming long enough, you already know who he is.
What you might not know is that he spent the last several years building AI tools to make gaming safer.
That’s GGWP. And the founder story is just the start of what makes it interesting.
What GGWP Actually Is
GGWP is a community safety platform for games, but calling it a moderation tool undersells what it does.
Where ToxMod (which we covered in depth here) was built around voice chat analysis, GGWP is built around the full picture of how toxic behavior shows up in multiplayer games. Voice is part of it. Text is part of it. But so is what players do, not just what they say.
The platform covers:
Text moderation across 18+ languages. Real-time analysis of chat that goes beyond keyword matching to look at context, intent, and conversational patterns.
Voice moderation across 12 languages. Similar to ToxMod in function, real-time audio analysis with flagging for human review, but built as one component of a broader system rather than the core product.
Username screening. Offensive, impersonating, or identity-targeting usernames caught before they ever appear in your lobby.
Player reputation profiles. This is one of the features that makes GGWP genuinely different. The system tracks behavior over time and across sessions, building a behavioral record for each player. Not just “did they say something bad today” but “what has this player’s behavior looked like across the past three months.” Someone who’s had a clean record for six months and has one bad session gets treated differently than someone with a pattern of violations.
Report triage with credibility scoring. GGWP tracks the reliability of the players filing reports. Someone who consistently files accurate reports gets more weight. Someone who repeatedly files false reports, reporting players who aren’t actually violating anything, gets less. This addresses a genuine abuse vector in most reporting systems, where bad actors can weaponize reports against players they simply don’t like.
In-game sentiment analysis. Beyond individual violations, GGWP tracks community health trends at the game level. Are certain maps producing more toxicity? Are certain game modes correlated with more identity-based harassment? That aggregate data goes back to developers who can use it to make design decisions.
Discord integration. GGWP connects with game servers on Discord, extending community safety coverage beyond the game itself to where communities actually live between sessions.
The pitch GGWP makes to developers is that piecemeal solutions, a voice tool here, a text filter there, create gaps and inconsistencies. A player banned in voice chat can still be a nightmare in text. A player with a clean record in the game has no accountability for what they do on the game’s Discord server. GGWP’s argument is that community safety needs to be a complete system, not a collection of individual tools.
The People Behind It
Fong co-founded GGWP with Kun Gao in 2020. If Fong’s name is the one that opens doors with gaming people, Gao’s background is what makes the business credible to investors.
Gao co-founded Crunchyroll, the anime streaming platform that eventually sold to Sony for over a billion dollars. He knows how to build a media platform, how to scale a consumer product, and how to navigate the content licensing and trust-and-safety issues that come with running a large online community. That combination of a gaming legend and a proven operator is part of why GGWP raised what it did.
The funding round was $33.3 million, and the investor list is worth reading carefully. Riot Games put money in. Sony Innovation Fund put money in. Samsung Ventures put money in. YouTube co-founder Steve Chen put money in.
Riot putting money into a community safety company is particularly notable. Riot has spent years building its own internal behavioral systems for League of Legends and Valorant, they arguably have the most sophisticated in-house moderation research of any game studio. The fact that they invested in GGWP rather than building a competing external product suggests they see the external market as worth supporting, not threatening.
Sony’s involvement connects to PlayStation’s broader trust-and-safety push. Microsoft built Community Sift into Xbox. Sony backed GGWP. The platform competition around online safety infrastructure is real, and it’s happening at the investment level, not just the product level.
The Unity Vivox Deal and Why It Matters
In March 2025, Unity selected GGWP as the official safety partner for Vivox, Unity’s voice chat middleware that powers thousands of multiplayer games.
This is the deal that changes GGWP’s trajectory.
Vivox is the voice infrastructure that runs in an enormous number of games, including plenty of titles you’d recognize. It’s not the most visible product in gaming, but it’s in the pipes of a massive portion of multiplayer game voice communication. Smaller studios that need voice chat capability license Vivox rather than build their own systems.
Before the GGWP partnership, those studios had no real path to voice moderation. ToxMod is priced for large studios with real volume. Hiring a moderation team isn’t realistic for an indie developer with 50,000 players. The options were a keyword filter or nothing.
GGWP has a free tier. And it’s now the safety partner for the voice middleware those indie developers are already using.
What this means in practice: the games that couldn’t afford moderation now have a path to it. An indie MOBA with a small but passionate community can now have voice moderation that would have been completely out of reach two years ago. The accessibility shift is significant, not just for those specific games, but for players who spend time in smaller multiplayer titles that never made the enterprise moderation shortlist.
The Case Study: What Happened in Predecessor
The most detailed public data GGWP has published comes from their work with Omeda Studios on Predecessor, a third-person MOBA that launched in early access and built a passionate community.
MOBAs are worth paying attention to from a toxicity standpoint because the genre has a historical problem with it. League of Legends built its entire behavioral systems operation partly in response to what became industry-notorious levels of player-on-player hostility. The competitive structure, the team-dependency, the long match lengths, the high skill floor, all of it creates conditions where things go bad and people make each other miserable.
Omeda deployed GGWP’s full platform and published results after the rollout:
56.3% reduction in offensive messages. More than half the offensive text content in the game disappeared. That’s a dramatic change to the actual experience of playing.
58% drop in identity-based incidents. The harassment targeting players based on race, gender, sexuality, or other identity markers, the stuff that doesn’t just make a session unpleasant but makes players feel genuinely unwelcome in the game, dropped by more than half.
30 times more incidents detected than manual moderation alone. This one puts the coverage question into perspective. Whatever human moderation team Omeda had before GGWP, it was catching roughly 3% of what the AI system could cover. Not because the human team was bad at their jobs, because the human team is physically limited and the AI isn’t.
These numbers line up with the ToxMod data from Call of Duty in an important way: they both show that AI moderation isn’t just catching the extreme obvious violations that would get flagged anyway. It’s catching the mid-level toxicity that used to slip through constantly. And that mid-level toxicity is, for most players, most of the problem.
How GGWP Compares to ToxMod
Since we covered ToxMod in the previous article, it’s worth being direct about how the two products compare. They’re not really competitors in the traditional sense, they serve overlapping but different markets, but if you’re a developer choosing between them, or a player trying to understand which system is running in your game, the differences matter.
ToxMod is voice-first. GGWP is community-first. ToxMod was built from the ground up as a voice analysis tool, and it’s the best in the industry at exactly that. GGWP treats voice as one channel among several and builds the broader community health picture. If your primary problem is voice chat harassment, ToxMod has more depth there. If your primary problem is toxicity across every communication channel and you want unified player reputation tracking, GGWP has more breadth.
ToxMod has more published voice-specific data. The Call of Duty transparency reports give ToxMod the most publicly documented track record of any moderation tool in gaming. GGWP’s Predecessor case study is solid, but ToxMod has two years of progress reports from one of the biggest game franchises in the world.
GGWP is more accessible to smaller developers. ToxMod is enterprise-priced, which is fine for Activision and Rockstar. GGWP has a free tier and a Unity Vivox integration that makes it reachable for indie studios. If you play smaller multiplayer games, GGWP is more likely to be the system running in them.
Both use human review for enforcement. Neither system auto-bans. Both flag for human review. This is a design principle the industry has broadly converged on, AI catches the volume that humans can’t cover, humans make the calls that require judgment. That’s the right approach.
The reputation tracking is GGWP’s real differentiator. ToxMod analyzes individual sessions. GGWP builds longitudinal player profiles. For catching people who fly under the threshold in any individual session but are clearly running a pattern of targeted harassment over time, the profile approach catches things session-by-session analysis misses.
The “Good Behavior” Side of the Equation
One thing worth highlighting that separates GGWP’s philosophy from a pure moderation tool: they’re explicitly focused on rewarding positive behavior, not just punishing negative behavior.
The player reputation profiles work in both directions. Positive behavior, helpful communication, consistent clean play, commendations from other players, builds a positive reputation that can affect matchmaking and feature access in games that implement the full system. It’s not just a record of your violations. It’s a record of your character as a community member.
Fong has talked about this publicly. His argument is that pure punishment systems have a ceiling. You can remove bad actors. You can discourage the worst behavior. But you can’t build a community you’d actually want to spend time in just by removing people. You have to give players reasons to behave well, not just consequences for behaving badly.
Valorant‘s Honor system works on similar logic, it rewards positive behavior with cosmetic unlocks and matchmaking benefits. Overwatch 2 has commendations. Deep Rock Galactic‘s entire culture runs on this principle structurally, even without formal reward systems.
GGWP’s version formalizes it into the AI infrastructure. The system tracks the carrots, not just the sticks.
What This Means for Players
GGWP is running in 25+ games as of the latest public numbers. Most of those aren’t household names, they’re mid-size multiplayer titles, indie competitive games, smaller communities that wouldn’t have had moderation infrastructure otherwise.
If you play in those spaces, GGWP is probably protecting your experience more than you realize. The Predecessor numbers aren’t an outlier. Games that deploy real moderation tools see real changes in community behavior. The question has always been whether those tools would ever be accessible to games outside the top tier of the industry.
The Unity Vivox deal answers that question more optimistically than I would have expected two years ago. A game doesn’t have to be Call of Duty to have AI moderation anymore. An indie studio with a few hundred thousand players can turn on GGWP through their existing voice infrastructure and see the same category of improvements that Activision is publishing reports about.
That’s genuinely good news for the parts of gaming that don’t make the front page of gaming publications but are where plenty of adults with limited time actually spend their gaming hours.
Play any games that use GGWP? Noticed a difference in your community after a moderation rollout? I’d genuinely like to know, drop it in the comments or find me in the TAG Discord. Smaller games with positive moderation stories are exactly what the TAG Community Safety Score is being built to recognize.
