How Rainbow Six® Siege Took Back Its Voice Chat with ToxMod to See a 50% Drop in Critical Toxic Voice Draft

Executive Summary

In 2025, Ubisoft made Rainbow Six Siege a free access title. The anticipated wave of new players garnered excitement among the team, but the switch to free access also opened the door to an influx of potentially toxic behavior no longer gated by a premium price tag. At the time, the team was looking for a way to better review voice chat or moderate content at-scale. Reports came in but nothing could be verified or actioned without extensive manual work.

Looking for the right solution, the Siege team kicked off a competitive evaluation process and a four-week trial with Modulate’s ToxMod to characterize the in-game environment and tune detection thresholds for their voice chat content moderation needs. Post-trial, Ubisoft and the Siege team decided ToxMod was the best solution to solve their challenges and went live in production in September of 2025. The impact was immediate and measurable.

75%

Decrease in per-session toxicity rate within two weeks of go-live (from 37% to 8%).

50%

Reduction in critical toxic voice chat events with a 30% decrease in general voice toxicity after in-game enforcement went live

800K hours

Of voice audio analyzed every month.

Up to 8% → 0%

Voice-chat toxicity was in a constant range from 3-8% of weekly community discussion over five years, then disappeared within three months of ToxMod going live.

The data tells a story beyond detection alone. In the weeks following go-live, both the toxicity rate and absolute incident volume declined quickly, meaning there was strong evidence in the data that player behavior actually changed. The additional visibility also provided intelligence that wasn’t possible before: ToxMod surfaced 9.5 million high-severity incidents across the first eight months, revealing that on any given day, roughly 35% of flagged users drive 66% of all incidents — insights that now help shape Ubisoft’s overall enforcement strategy.

Don’t think of voice moderation only as a moderation project — it’s a player experience investment. When policy, operations, production, and community expectations come together, you’re not just reducing harmful behavior. You’re building healthier communities and stronger trust in your games.”
‍
– Jeremy Marchadier, Senior Director, Global Player Safety, Ubisoft

Background

The Most Competitive Ubisoft Title (and hardest to moderate)

Siege occupies a unique position at Ubisoft: it’s the company’s flagship online competitive first-person shooter, a ten-year-old title, and home to a fiercely dedicated player base. And by design, it’s intense. When you die in Siege, you don’t respawn like in other games. You watch from the sidelines. Friendly fire and team kills are possible, every round is high-stakes, and some amount of frustration is baked into the risk/reward experience.

Voice chat sits at the center of all of it. Siege provides open-mic team voice for every player in every match. It’s not proximity-based chat that fades in and out like other shooters based on how close you are to teammates. If you’re using voice chat in Siege, you likely know the game well: the callouts, the maps, the meta progression of weapons and operators. Communication isn’t simply a social garnish. It’s a way to create a competitive advantage.

The combination of a high-pressure competitive format plus always-on, full-team, voice chat makes the likelihood of toxicity meaningfully higher in Siege than in most other noncompetitive games. The Siege team knew this was a factor to consider, however, they needed a way to see it.

“The nature of the game makes it competitive. We don’t have the same challenges as other popular competitive first person shooters. We have an open-mic voice chat for everyone on your team, every time. The likelihood of toxic behavior is way higher.”
‍
– Lancelot Satge, Game Protection Director, Rainbow Six Siege, Ubisoft

The Problem

A Decade of Voice Chat Without Sufficient Visibility

Lancelot Satge, Rainbow Six Siege’s Game Protection Director, said this about how the team handled toxic voice behavior before ToxMod:

“We couldn’t really do much. We weren’t saving recordings. Even if a player was reported for voice chat, we couldn’t validate. Before ToxMod, we didn’t have a viable solution.”
‍
– Lancelot Satge, Game Protection Director, Rainbow Six Siege, Ubisoft

Text chat could be filtered and reports could be filed. But voice chat was more akin to the wild west. Ubisoft’s broader approach relied on manual processes distributed across multiple teams, which preserved human judgment but couldn’t deliver the responsiveness a live-service game demands on a continuous basis. Voice made it harder because it’s inherently contextual, nuanced, and impossible to interpret at scale without dedicated tooling like ToxMod.

For the gaming industry as a whole, it’s no secret that toxic interactions can degrade the quality of play and impact player wellbeing. For a game as popular as Siege, this reality was unavoidable and something the team took seriously. In Ubisoft’s weekly community sentiment reports, which screen social media for the most popular player topics, voice chat toxicity appeared in 2–4% of all discussions for five years straight.

By 2024, growing report volumes and the lack of clear visibility made automated voice moderation a strategic priority. Then came the decision to open up Rainbow Six Siege as a free access title, creating a sense of urgency to find a voice moderation solution. More players were coming and with that, the potential for more toxicity. It was critical for Ubisoft and the Siege team to be ready before it arrived.