I think the biggest challenge is to create players who can act like humans would," Fabio Zinno, senior software engineer at Ea,
You can train an autonomous agent to play Fifa and win every time, but if its movements are unnatural, erratic and unpredictable, you probably failed to create a good game AI." The goal shouldn't be to win at all costs, but to create a compelling, challenging and overall fun experience for our players.
We are very committed to creating smarter tools for content creation, including 3D models and textures, terrain, stadiums, cityscapes, facial and body animations. A lot of effort is going into motion capture of facial and body movements from video, a technology that will not only give us a higher volume of animations, but also much more organic and natural results: as good as the animations captured with the traditional studio motion capture process are, they are done using markers and props that can sometimes hinder natural movement. Being able to capture the movements of athletes "in their natural environment," especially in sports titles, is a very tempting goal. Machine Learning is also enabling more holistic solutions for character performance, and we're working on models that can generate speech from text, along with lip sync and facial animations. The goal is to include body gestures in the future to provide realistic and convincing behaviors.
Performance analysis of real athletes and teams in competitive sports is a hot topic, and not just for sports game developers: sports like soccer, football, and field hockey are increasingly implementing hardware and software solutions to capture player and team performance, extract relevant metrics, data, and foster a more efficient and scientific approach to training and strategy. Such an approach would allow us to create game AI that can more accurately reproduce the behavior of real teams and world-class athletes. This data could be provided by the leagues themselves (NFL, NHL, FIFA), but in the future I expect we will have technology mature enough to work directly with footage from television broadcasts: imagine being able to recreate the Champions League final and offer it as new content for FIFA players the following week!