Machine learning is changing the $1 billion esports market, which is projected to reach $1.8 billion by 2022. Computer scientists have made great strides over the past decade to vastly improve the way AI learns human language, big data, and strategy. AI in gaming is rapidly changing, including player-athlete performance, conversational assistants, game design, and discovery of new approaches to game theory and in-game strategies.
Optimizing esports performance
Professional gamers have rigorous routines, just like NBA and NFL players. These teens and twentysomethings compete intensely for six- and seven-figure salaries by way of tournament prizes that can propel them to stardom. They’ve also got the dough to pay for AI tools as business expense.
For example, esports analytics platforms such as our SenpAI are providing AI-powered coaching that can assess player stats, and suggest better strategies in MOBA games like for League of Legends and Dota 2. Each game is played by two teams who must each defend a base. An AI coach advises team members on how to attack and defend, and shows how alternative approaches can increase (or lessen) the odds of winning.
Developers train AI agents to learn specific games. In the case of Omnicoach, Overwatch players get useful tips on how to use weapons, improve mobility, and secure favorable positions against enemy avatars that each possess unique combat-fighting capabilities. In this game, teams of six members who play with great synergy enjoy an edge over opponents.
Elite gamers (who are backed by global brands like Red Bull, Monster Energy, and Audi) are adopting computer-generated game plans to gain a tactical edge. In Counter-Strike: Global Offensive, for example, an AI coach can teach the player to hide behind favorable positions near a building or bridge to better shoot opposing players who rush through openings, or to deploy team formations that make them less vulnerable to a counter-attack.
Simulated warfare and game theory are influenced by plenty of variables. AI brings analytical horsepower that’s helpful to recreational folks. But it’s invaluable to pros who are serious at winning because their livelihoods are at stake.
Devising winning game plans
Artificial intelligence is changing video gaming in other ways. London-based DeepMind (acquired by Google in 2014) used machine-learning to discover better ways of beating old-school games such as Pong by Atari, and other staples at your local arcade or movie theater.
In a 2017 TedTalk, DeepMind computer scientist Raia Hadsell said that AI and deep neural networks can solve games that we play, but also improve game design. According to Hadsell, games are the ultimate test lab for AI because we can observe the results –that is, gaming performance is not subjective.
AI tech has advanced to a point where it’s now virtually impossible for humans (even chess legend Gary Kasparov) to defeat supercomputers in chess, checkers, and other games. Indeed, we have entered an age where reasonably-priced machines can calculate the consequences of millions of in-game maneuvers per second.
There’s less distinction between what’s virtual and what’s physical, anymore. Pro gamers are immersing themselves in a blended universe of actual and digital—of human and artificial allies and opponents.
Disrupting gaming’s business models
Monetization is affected as well. Conversational AI assistants are being developed to help consumers navigate through a labyrinthian maze of thousands of video-game titles. And to find games that match consumer tastes. Soon, all you’d have to do is speak to your iPhone and ask a lifelike avatar what video games are on sale; or which products are best sellers; or a host of other queries that save time and money.
“Customers demand a natural, conversational interaction when shopping online that’s similar to an in-store experience,” says Vijay Ramakrishnan, a Silicon Valley-based machine-learning engineer who has developed AI assistants. “People want AI to recommend products instead of consumers finding items themselves.”
Moreover, AI portends future use of capturing market share by creating game designs that vastly increase player interest and engagement. For example, an algorithmic program may be trained to find the best in-game features of Dota 2 and League of Legends and other popular titles to help MOBA developers hone their products.
Mobile and social platforms
“Online chatbot channels like Facebook Messenger are great at delivering high-quality AI experiences,” says Ramakrishnan. “Facebook can peruse past interactions and behavior with an AI agent for ideas on what a person may want next—such as “liked” items and favorite places.” These insights can lead to offers that uniquely resonates with an individual.
Thus, machine-learning is suited for customizing games based on user preferences. AI agents can learn to guess exactly what an individual will likely purchase. And this capability reduces the financial risk for game developers. Because there’s no value in designing experiences that a supercomputer “knows” (with high probability) that the marketplace will reject as unimaginative or uninspiring.
“Multi-modal AI that combines an object and face-recognition system, a voice assistant, and display screen can provide an immersive experience,” says Ramakrishnan.
Russian President Vladimir Putin said in 2017 that “Artificial intelligence is the future … Whoever becomes the leader in this sphere will become the ruler of the world.” And perhaps ruler of esports domain.
Berk Ozer is a co-founder of FalconAI, a VC-backed startup that develops cutting-edge AI algorithms to democratize domain knowledge of human experts.
You can't solo security COVID-19 game security report: Learn the latest attack trends in gaming. Access here