EA could make its UFC and Fight Night games better with AI

With the technology, EA can develop more lifelike character animations for use in its video games, especially martial arts titles.


EA's patent reveals a machine learning technology that will improve martial arts gameplay.
EA's patent reveals a machine learning technology that will improve martial arts gameplay.

The AI age is here and Electronic Arts, like many other companies, including Disney, which we recently reported may actually be looking to buy EA, is taking advantage of the technology to improve its services.

According to a recently published patent, EA is developing a machine learning system that will dynamically generate improved and more life-like animations for its video game characters. The system will rely on data from motion capture video to achieve this goal.

EA originally filed this patent in 2021, so there’s no telling how much progress it has made since. If it proves viable, EA developers will be able to generate more realistic motion for video game characters with less work than ever before and in less time, using unstructured motion capture data.

EA’s new technology, geared specifically towards its fighting games, will blend different bodily motions to create life-like transitions between various attacks, blocks, counters, and more. The system will use a modular neural network architecture – a problem-solving technology – to separately model diverse movement scenarios from raw motion capture data.

Watch this story and subscribe to Xfire on YouTube.

However, the technology will place the ultimate power for determining what is rendered on the screen during gameplay with the user, via interchangeable controllers.

Consequently, EA’s developing technology will enable martial arts games to depict a diverse range of martial arts skills and signature moves similar to those of real-life fighters. Furthermore, it will provide gamers with the ability to combine these different moves dynamically and even create new and original moves with no real-world reference. Gamers will be able to move fluidly between punching their opponents and kicking them, as well as blocking moves and countering instantly, among other things.

EA's machine learning system should also have a use case for sports video games.
EA’s machine learning system should also have a use case for sports video games.

Popular for fighting game franchises like UFC, EA Sports MMA, and Fight Night, this technology will give EA an edge in action video game franchises too, if it is successful.

Recently, EA decided to move away from the FIFA brand for its soccer video game franchise – a decision that is already paying dividends. Its latest installment in that series, EA Sports FC 24, received generally favorable reviews from critics and reached over 11 million players in its first week.

If the new technology is successful, it will help the company further extend its dominance in the sports video game category as well, including the aforementioned EA Sports FC franchise and the Madden NFL franchise, as those games also require a diverse range of movements.

EA’s recently revealed system is one of several machine learning-based technologies in development at the company. According to reports, the company is also working on two other systems, Project Meade and Machine Learning Trees, that also seek to improve the gameplay experience for gamers.

Despite the many potential benefits, EA’s machine learning technology isn’t without its challenges. For one, it needs to grapple with ensuring user privacy as it develops these systems, especially since it requires the use of biometrics data.

If EA manages to pull it off, then it will have made a breakthrough that could put it right at the forefront of a revolution in the gaming industry, ultimately delivering an incredible gaming experience to its customers and robust profit to its investors.

Tobi Oguntola
Tobi Oguntola // Articles: 777