Unity has found a way to reduce toxicity in video games by using AI.

Unity has found a way to reduce toxicity in video games by using AI. (Source – Shutterstock)

Tired of the toxicity in video games? Unity is solving that problem with AI

  • Unity’s “Safe Voice” uses AI and sound intelligence to tackle online gaming toxicity, with potential applications beyond gaming.
  • Positive gaming communities boost revenue; Unity’s Safe Voice proactively identifies and curbs toxic behaviors.
  • “Safe Voice” technology has potential for calming online interaction beyond the gaming world.

In the digital age, online gaming has transcended mere entertainment, morphing into vibrant virtual societies that mirror the complexity and diversity of the real world. With the use of AI innovations from Unity on the horizon, these platforms, where barriers dissolve and cultures intersect, are not just facilitating unprecedented social engagement and camaraderie, but are also paving the way to tackle inherent challenges like toxicity.

Yet, alongside these positive aspects, toxicity and miscommunication still lurk, which can detract from the immersive experience gamers are looking for. In a recent study conducted by Dr. Rachel Kowert, the research director of Take This, it was revealed that toxic game communities are detrimental to the mental health of the wider gaming community and the game’s bottom line — its commercial and financial success.

The report highlights the connection between highly toxic gaming communities and their negative impact on a game’s success, with the ultimate goal of encouraging publishers to take stricter actions to combat toxicity in online games.

Dr. Kowert points out that games known for fostering a positive, safe community experienced a 54% gain in revenue compared to games generally regarded as having toxic, abusive communities. With the aid of Nielsen, a global leader in data and analytics, it was further revealed that players were much more likely to quit playing any game if faced with toxic and abusive behavior.

Unity presents “Safe Voice” – The AI and sound intelligence

Ensuring a safe and harmonious environment is therefore crucial for the evolving world of online gaming. Technologies are emerging to meet this demand, with innovations like Unity’s recent offering poised to herald a new chapter in reimagining and elevating the virtual gaming ecosystem.

Tech Wire Asia reached out to Jeff Collins, SVP of Engineering, Operate Solutions, Unity, to discuss the technology behind Safe Voice, how it uses acoustic and semantic intelligence, and how it employs machine learning algorithms to identify toxic behavior and respond to player-initiated reports proactively.

Collins said that Safe Voice uses sophisticated machine learning techniques to analyze the context of voice communications in-game. Unlike traditional speech recognition technology, it employs acoustic and semantic analysis, focusing on intent and interaction quality rather than just words, providing a richer dataset, including emotions, satisfaction levels, and vocal bursts.

Jeff Collins, SVP of Engineering, Operate Solutions, Unity, explains how AI is being used in its new innovation.

Jeff Collins, SVP of Engineering, Operate Solutions, Unity, explains how AI is being used in its new innovation.

“Safe Voice kicks in when players flag instances of toxic or disruptive behaviors and developers can also proactively monitor activities in their game, like player mutes or abrupt exits,” he added.

Unity’s Safe Voice technology represents a sophisticated fusion of machine learning, acoustic intelligence, and semantic understanding. By using intricate algorithms, Safe Voice can analyze and interpret voice communications beyond transcription. This system discerns subtle tonal variations, from laughter to sighs, and evaluates the context behind spoken words, ensuring accuracy and reducing misinterpretations.

Deep dive into acoustic and semantic intelligence

Although we didn’t get the sense of the exact programming language behind Safe Voice, given Unity’s longstanding affinity for C#, it’s plausible that a substantial portion of the solution utilizes this language. However, given the system’s complex AI and natural language processing functions, it wouldn’t be surprising if Python, favored for its data science prowess, also played a role, possibly incorporating frameworks like TensorFlow or PyTorch.

With real-time monitoring capabilities, Safe Voice also underscores the need for efficient data processing, potentially using distributed computing. The culmination of these technologies not only showcases Unity’s forward-thinking approach but also emphasizes the evolving synergy of gaming and advanced computational methodologies.

For those unfamiliar with acoustic and semantic intelligence, in the context of voice analysis, acoustic intelligence refers to the understanding and interpretation of sound patterns. This doesn’t just focus on the spoken words, but also the non-verbal aspects of speech. It includes factors such as tone, pitch, loudness, and vocal bursts (like laughter or sighs).

Acoustic intelligence can pick up on the emotional undertones of a message, capturing whether someone is angry, joyful, sarcastic, stressed, etc. For instance, the same sentence spoken in different tones can convey happiness, sarcasm, or anger.

Deep dive into coding with Unity in creating Safe Voice - with the inclusion of AI.

Deep dive into coding with Unity in creating Safe Voice. (Source – Shutterstock)

On the other hand, semantic intelligence delves deeper into the meaning and context of the words spoken. It’s not just about the words used, but how they’re used with each other, the intent behind them, and the meaning they convey in specific contexts.

This intelligence goes beyond basic speech recognition (which transcribes words) to understand the message’s content, context, and intent. For example, the word “bank” can mean a financial institution or the side of a river, and semantic intelligence would determine the intended meaning based on context.

So, using both acoustic and semantic intelligence for Safe Voice allows for a comprehensive understanding of in-game voice interactions. While acoustic intelligence can identify potential emotional outbursts or heated exchanges through vocal tone and volume, semantic intelligence ensures that the software doesn’t misinterpret a word or phrase that might be innocuous in context but flagged as potentially toxic when taken out of context.

The combination ensures both the emotional tone and the actual content of the conversation are analyzed for a more accurate picture of the in-game interaction. Safe Voice’s detailed session reporting dashboards give teams a thorough and nuanced understanding of player interactions.

From detection to action – AI usage with Unity

According to Collins, the overview dashboard shows trends over time and surfaces behaviors driving the most disruption in their game. Developers can see a detailed breakdown of player attributes, behavior, and categorize sessions as toxic or vulnerable and have visibility into the status of player consent.

“The monitoring dashboard is updated in near-real time and reflects their custom priorities so they can focus their efforts,” said Collins. “With Safe Voice, moderators are empowered to make informed decisions promptly and efficiently, allowing them to mitigate situations before they have a chance to escalate.”

Determining toxic behavior on a platform needs that combination approach, because obviously, when talking about toxicity, we’re talking about a mindset expressed through analysis of player behaviors like tone, loudness, pitch, and emotion. According to Collins, Safe Voice can offer a deeper understanding of player behaviors through analyzing intent, which is the most powerful indicator of interaction quality.

The solution captures a rich dataset, including vocal attributes such as tone, loudness, pitch, and emotion. By continuously monitoring and assessing these characteristics, AI can recognize patterns corresponding to specific in-game actions, player states, or engagement levels to help studios moderate toxicity in real-time.

As a result, Safe Voice helps create a more positive gaming environment for players by allowing moderators to more effectively and efficiently take action against those who are being toxic. Additionally, it can even be used to reward supportive players. By equipping studios with more insights about toxic behavior within their communities, it helps inform how they adapt their current game and can also help guide the development of future games, so they are more safe by design.

Actions taken when toxicity is taken place in Xbox - Unity has an answer with the use of AI

Actions taken when toxicity is taken place in Xbox. (Source – X)

Questions then arise on the impact on gaming communities: what changes might players see? How could this influence the culture of online gaming as a whole?

With the implementation of Safe Voice, players may notice a significant decrease in toxic behavior within online gaming communities. Collins breaks it down by saying that a cleaner and more respectful communication environment could foster a stronger sense of safety and inclusivity, encouraging diverse participation and cooperation among players. This can lead to a richer and more enjoyable gaming experience, as individuals feel more comfortable engaging in teamwork, sharing ideas, and collaborating on strategies.

Simultaneously, this shift in communication culture could have broader impacts on the gaming community. The standards for acceptable behavior may rise, creating a more welcoming and positive environment, which could result in a cultural shift towards respect and empathy.

Collins emphasized that “Our hope is to be a catalyst for transformation towards a less toxic and more engaging gaming culture, possibly inspiring developers to create games that further encourage positive interaction.”

Going beyond Unity gaming: The wider promise of AI

Seeing the potential of this technology, it would be beneficial to extend it outside of gaming, for instance, to social media and other online communities. Imagine a digital ecosystem where platforms such as social media, discussion forums, and even professional collaboration tools incorporate Safe Voice-like technologies.

Social media platforms, which are often rife with heated exchanges, could greatly benefit from real-time monitoring and moderation. Such a system could identify potential flare-ups before they escalate, steering conversations back to civility or intervening when necessary. This has the potential to reduce online bullying, hate speech, and misinformation significantly.

As AI continues to evolve, understanding the nuances of human interaction becomes paramount. While today it aids in identifying toxic behavior in games, tomorrow it might be employed in customer service to gauge satisfaction, in telemedicine to discern patients’ emotional well-being, or in online education to understand students’ levels of engagement.

Unfortunately, for now, Unity is focused on gaming. The toxicity problem in gaming has been well documented, and there has been a demand from both developers and players to address this challenge.

“We’re very excited to offer a solution to help creators foster more positive gaming environments. We believe the benefits of Safe Voice can be applied to other areas beyond gaming as well, and it will be something we explore in the future,” Collins concluded.

Dudebros, get a grip – or a grip will be gotten for you.