3 Games Where Emerging Tech Is Beating Humans – And What We Can Learn From Them

Digital transformation—and emerging technologies like artificial intelligence (AI)—have paved the way for many solutions to some of the world’s most complex problems. In fact, Statista reports that digitally-transformed organizations will contribute more than half of the world’s GDP by 2023.

But before novel technologies can find real-life applications, AI researchers use games to improve their efficacy. AI has been a feature of games for many years, with top titles such as the cultural phenomenon Grand Theft Auto obvious examples. In that game, the actions of NPCs, non-player characters, are controlled by AI to give the user a challenging experience. If you play sports video games, often you’ll pit your wits against the AI. If you lose, it’s not a surprise, but what if you were in real life? What if you played a real game, a physical sport or hobby, and were beaten by AI?

Here are some games where emerging tech is beating humans—and what we can learn from them.

Jeopardy!

Because of how random the questions are each night, the best Jeopardy! contestants are those who have a broad knowledge of a variety of topics. This is how prodigy Ken Jennings, one of the show’s most memorable contestants, won a streak of 74 Jeopardy! games.

In 2009, IBM partnered with him to develop a supercomputer, the IBM Watson, which learned by playing against the world’s top Jeopardy! players. In 2011, Watson beat Jennings and 20-time Jeopardy! winner Brad Rutter in a three-day contest. This success exhibited the power of artificial intelligence to surpass even the best of human capacities. Following Watson’s win, other tech giants have begun to jump in on deep learning and AI research.

Poker

AI has shown that it can turn to methods alternative to the ones humans commonly use to win. This is most evident in AI winning against humans in poker games. Such a feat is notable considering that playing Texas Hold’em requires focusing on other players and having a deep understanding of human psychology to determine their patterns and risk-taking qualities.

AI program Pluribus beat some of the world’s best poker players in a six-player round of Texas Hold’em. Usually, an AI wins a game by generating decision trees throughout the game. But AI learns through reinforcement learning, which involves evaluating previous games and assessing success based on the circumstances it’s presented with. Before playing with humans, Pluribus played trillions of hands on its own. This helped it learn how to narrow down its decision trees to only a few moves ahead, as more opportunities for change are present in a game with six players.

Chess

AlphaZero is an AI program that stunned scientists worldwide when it was able to beat the world’s leading chess program, Stockfish, after just four hours of playing on its own. Its triumph has given rise to its incorporation into quantum computing.

Quantum mechanics has come a long way since the discovery of the Planck equation, which broadened our knowledge of calculating a quantum system’s temporal evolution of energy. Now, these principles are used in quantum computing technology. Where supercomputers fail, quantum computers create multidimensional spaces where patterns that link individual data points emerge.

When a research group from Aarhus University recognized the broad applicability of AlphaZero following its chess triumph, they reviewed the program’s data and learned that it had harnessed data from an underlying symmetry problem they had not originally considered. Realizing that AlphaZero worked best when combined with a specialized quantum optimization algorithm, they made their code open-source to speed up progress in the field.

In each of the above instances, computers solved problems by approaching them in ways humans can’t. But it’s integral to note that although emerging technology can be applied to provide solutions to some of the world’s most critical issues like cybersecurity and fraud detection, humans still have to be present to improve them. As the Aarhus research group gleaned, emerging technology’s next goal will be to develop hybrid intelligence interfaces that leverage and combine the strengths of humans and computers alike.

Leave a Reply

Your email address will not be published. Required fields are marked *