PUBG: AI Can Guide You, What Is Right & Wrong In The Game

Krafton has shared that he checks some 100,000 accounts weekly, a huge effort that requires more resources. 

Creating good online communities is not an easy task. The great multiplayer successes also present the challenge of visibility and the huge number of players they manage. Therefore, Krafton not only boasts about the number of accounts he bans every week in BGMI, but also delegates much of the work to the AI.

The Chinese company, far from wanting to lose ground in terms of battle royale, has announced on its official blog that it is expelling approximately 100,000 accounts from the game, although the average is “between 60,000 and 100,000 players.” An absurdly large number that Krafton assumes are accounts involved in the use, distribution, or sale of illegal software.

If we look at the number of players that this battle royale has on Steam, the only platform with which it can be measured, we see that these 100,000 users correspond to a third of the users on PC. A spectacular number that has led the Chinese team, especially considering the work that this entails, to ” a more comprehensive approach”.

BGMI with AI Mode

Also Read: BGMI Server Authentication Error: How To Fix It? Easy Steps & Process

To accomplish this, Krafton is looking for a ” fundamental solution ” that allows accounts that cheat in competitive mode to be analyzed and tracked. Its focus is on two types of accounts: “hijacked accounts” and accounts that “exploit the survival mastery level system.”

According to Krafton, its analysis shows that approximately 85% of permanently locked accounts were created before the transition to the free-to-play model. This “implies that cheaters have most likely obtained other players’ accounts to start using illegal software on those accounts.”

With this in mind, the company has worked on a ” machine learning model that could learn the characteristics and patterns of domain-level abuse.” As they say, the Krafton team has been working on this since the beginning of the year, although it is not yet complete. The idea is that the development team continues to improve an AI that, for the moment, “has tripled the number of reported users compared to the period before the introduction of this model.”

Of course, facing the player this is a problem. Will the AI ​​be able to discern well what is happening depending on what situation? Here Krafton itself comes into play, it should not relegate the entire job to the self-learning AI if it does not want the situation to turn against it.