Carnegie Mellon's Poker-Playing AI Loses Tournament

Claudico, an artificial intelligence (AI) program created at Carnegie Mellon University finished in second-to-last place in the "Brains Vs. Artificial Intelligence" exhibition tournament that began on April 24 at Rivers Casino in Pennsylvania, but the difference in score is not large enough to be statistically significant, according to information from the university.

During the two-week tournament, Claudico played 20,000 hands of poker with each of four of the world's best players of Heads-Up No-Limit Texas Hold'em. "Bjorn Li had an individual chip total of $529,033, Doug Polk had $213,671 and Dong Kim had $70,491. Jason Les trailed Claudico by $80,482," stated a news release from Carnegie Mellon. No actual wagering took place during the tournament, but the four professional players earned appearance fees based on their performance.

While the results of the tournament were not statistically significant, they did provide Claudico's developers with a plethora of data to improve its algorithms. The scientists plan to use the 80,000 hands of data to train and test future iterations of the AI program.

One area for improvement is the program's betting logic. "Where a human might place a bet worth half or three-quarters of the pot, Claudico would sometimes bet a miserly 10 percent or an over-the-top 1,000 percent," stated the news release.

While the scientists are striving to develop software that can beat humans at poker, the end-goal of the endeavor is to develop artificial intelligence that can make good decisions in the face of incomplete information. Poker is an ideal testing ground for that type of AI because the game requires players to mislead their opponents. If it can learn to detect deceptions such as bluffing and show play, it can learn to help people solve problems.

"Beating humans isn't really our goal; it's just a milestone along the way," said Tuomas Sandholm, a computer science professor at Carnegie Mellon who has led development of Claudico, in a prepared statement. "What we want to do is create an artificial intelligence that can help humans negotiate or make decisions in situations where they can't know all of the facts."

Judging by the advances the scientists have made over the last eight months since Claudico's predecessor, Tartanian7, competed in a poker tournament, they believe their AI may be able to outwit the world's best players in as little as one year.

About the Author

Leila Meyer is a technology writer based in British Columbia. She can be reached at [email protected].

Featured

  •  laptop on a clean desk with digital padlock icon on the screen

    Study: Data Privacy a Top Concern as Orgs Scale Up AI Agents

    As organizations race to integrate AI agents into their cloud operations and business workflows, they face a crucial reality: while enthusiasm is high, major adoption barriers remain, according to a new Cloudera report. Chief among them is the challenge of safeguarding sensitive data.

  • glowing digital brain above a chessboard with data charts and flowcharts

    Why AI Strategy Matters (and Why Not Having One Is Risky)

    If your institution hasn't started developing an AI strategy, you are likely putting yourself and your stakeholders at risk, particularly when it comes to ethical use, responsible pedagogical and data practices, and innovative exploration.

  • college students in a classroom focus on a silver laptop, with a neural network diagram on the monitor in the background

    Report: 93% of Students Believe Gen AI Training Belongs in Degree Programs

    The vast majority of today's college students — 93% — believe generative AI training should be included in degree programs, according to a recent Coursera report. What's more, 86% of students consider gen AI the most crucial technical skill for career preparation, prioritizing it above in-demand skills such as data strategy and software development.

  • glowing blue AI sphere connected by fine light lines, positioned next to a red-orange shield with a checkmark

    Cloud Security Alliance Offers Playbook for Red Teaming Agentic AI Systems

    The Cloud Security Alliance has introduced a guide for red teaming Agentic AI systems, targeting the security and testing challenges posed by increasingly autonomous artificial intelligence.