Microsoft may soon use AI to scan for toxic words in Xbox party chats

The rise of AI is inevitable. Even a couple of months can make a massive difference in AI development. It’s been used in a lot of areas, and gaming is not immune to that. And now, it seems like Microsoft wants to use AI to scan for toxic words in Xbox party chats.

The Redmond tech giant recently shared its fourth Transparency Report, detailing how it can use the latest AI tech for gaming. In the announcement, Microsoft details its use of AI alongside human expertise to detect and address potential toxicity, complying with the Responsible AI Standard.

Some of the ways described are tools called Auto Labeling and Image Pattern Matching. The first assists in classifying conversational text by identifying potentially harmful words or phrases. Then, the latter enables rapid removal of known harmful content and identification of emergent toxic imagery using advanced database and image matching techniques.

A little while ago, the green console’s camp also introduced a new strike system. Players start with zero strikes and can get up to eight over six months, such as hate speech and profanity. Strikes lead to suspensions from social features, with severe violations possibly resulting in permanent suspensions

Microsoft, which owns Xbox, is under pressure to deliver more original content. Despite plenty of success with recent acquisitions of Activision and other game studios, Microsoft had to close a few Bethesda studios. 

An ex-Microsoft employee analyzes that the closure of several Bethesda studios may be linked to challenges faced by Game Pass and the recent ABK deal. 

Game Pass, once a boon for game exposure, is now facing revenue challenges due to too many subscriptions and changing player interests. And, with heavy investments in these acquisitions, Xbox is under pressure to deliver returns.