Roblox has acquired speech technology startup Speechly to help it integrate a new spatial voice chat feature safely. Roblox is a powerful sandbox tool that lets almost anyone create a unique game with its engine, letting the users do most of the work.

Many of these games attract an astonishing amount of players to the Roblox platform, totaling over 214 million monthly active users. Roblox has historically been a children’s game (and it still is) but its aging player base is causing its leadership to consider adding more mature games and features to keep them playing. Roblox’s fastest-growing user demographic in 2023 is 17-24 year-olds according to Demand Sage.

This is at least partially why the platform has added features like spatial voice chat along with more ambitious plans like building a new dating app for adult Roblox users. Many similar popular games within the same genre or similar veins like metaverse games have implemented these voice chat features so Roblox wanted to follow suit.

Management knows that it has to be careful with these additions to avoid putting its young users in harm’s way. Roblox conducted a study that found that 72% of gamers who have used voice chat at least once said that they have experienced a toxic incident. Immediately adding voice chat throughout its platform without sufficient moderation could be a total disaster.

This is where Speechly comes in. The startup was founded in 2016 with the mission to “enable more delightful computer voice interactions and to empower and enhance communication between people.” Its artificial intelligence-powered models can moderate speech in real time, making sure that Roblox’s users experience as little toxicity in voice chats as possible.

Is an Accurate Real-Time Voice Moderation System Even Possible?

This will not be an easy task. As voice chat is implemented into more and more experiences (what Roblox calls its user-created minigames), the sheer volume of speech will be extremely difficult to moderate effectively. Moderating text chats is relatively easy in comparison and only requires a simple set of banned words or phrases and detection methods to catch them.

Real-time voice chat moderation, on the other hand, requires the system to parse the speech first before it can tell if it has any banned words or phrases in it. This currently takes significant processing power to complete quickly, which is a vital requirement for real-time moderation. Roblox doesn’t want its users to have to wait seconds or even minutes for their voice chat to be heard by other players.

Roblox’s announcement of its acquisition of Speechly came less than 3 weeks after Activision revealed that it is implementing a similar voice moderation feature into its upcoming Call of Duty game: Call of Duty: Modern Warfare III. The game is supposed to come out on November 8th, 2023, and is set to be the gaming world’s first taste of real-time AI voice moderation.

If Activision is successful in building a strong moderation system by the game’s launch it will give credence to the idea that the technology can actually work. Roblox is lucky to be given the chance to study Activision’s successes and failures before it rolls out its own system with the help of Speechly.

Nevertheless, Roblox is still significantly more popular than the most recent Call of Duty title and may struggle even more with the sheer amount of data to parse from its 65.5 million daily active users.

Can Speechly Handle Roblox’s Voice Chat Moderation?

If anyone is up to the challenge it’s likely Speechly. It’s backed by Y-combinator and has a long list of other investors and supporters including Cherry Ventures, TQ Ventures, Seedcamp, and more. It has raised roughly $7 million in total from these and more investors and has been able to build up a suite of valuable IPs with these funds.

Speechly claims that its speech recognition technology is low latency, which is what makes real-time voice moderation possible. It actually starts responding to speech before the user is done speaking, unlike other competing systems. This could also allow Roblox to implement voice commands to make navigating the platform even easier than it currently is.

Voicebot.ai, the leading media platform covering the voice and conversational AI industry, argued that Speechly’s tech is ahead of the rest of the industry, including Modulate, the startup working with Activision to build its moderation system. It’s no surprise that Roblox chose the most advanced startup in the industry to help implement safe voice chat to its gargantuan user base.