In what could reshape the future of everything from QA testing to procedurally generated animation, EA's Search for Extraordinary Experiences Division (SEED) has built a self-learning AI-agent capable of teaching itself from scratch to play Battlefield I.
In recent years, self-learning agents like Google's DeepMind AI have learnt to play old Atari games and beat Go world champions, but this is the first time one has been able to self-learn something as complex as a modern AAA game.
The agent is currently proficient at basic Battlefield gameplay and after playtests, participants asked the developers to mark out the agents so they could be distinguished, which indicates a certain level of lifelike performance.
Although impressive the technology is still far from perfect; even after roughly 300 days of total gameplay experience, the AI was unable to devise or execute any strategies, and would sometimes find itself stuck running in circles. It is smart enough though to adapt its behaviour based on certain triggers such as health and ammunition.
"I'm confident they will do less silly stuff in the future, as they become more adept," said SEED technical director Magnus Nordin, adding that they while they are constantly improving, the process is slow.
The potential applications for this type of self-learning AI extends far beyond simply using it to compete against humans in multiplayer matches, however.
"Our short-term objective with this project has been to help the DICE team scale up its quality assurance and testing, which would help the studio to collect more crash reports and find more bugs," Nordin continued.
"In future titles, as deep learning technology matures, I expect self-learning agents to be part of the games themselves, as truly intelligent NPCs that can master a range of tasks, and that adapt and evolve over time as they accumulate experience from engaging with human players.
"I have no doubt in my mind that neural nets will start to gradually make their way into games in the years to come. Self-learning agents aren't just a good replacement for old-fashioned bots, you can also apply machine learning to a number of fields, such as procedurally generated content, animation, voice generation, speech recognition and more."