Game-playing AI

Programs that master games, AI's showcase benchmark from Samuel's self-learning checkers to AlphaGo's defeat of Go champion Lee Sedol.

AI & roboticsSoftware
  1. 1952

    First version

    Arthur Samuel (IBM)

    Image: Dan · CC BY-SA 2.0 · Wikimedia Commons
  2. 1997
    IBM Deep Blue

    IBM Deep Blue

    Defeats world chess champion Garry Kasparov in a six-game match.

    Image: Christina Xu · CC BY 2.0 · Wikimedia Commons
  3. 2011
    IBM Watson

    IBM Watson

    Wins Jeopardy! against Ken Jennings and Brad Rutter using open-domain question answering.

    Image: Clockready · CC BY-SA 3.0 · Wikimedia Commons
  4. 2016

    AlphaGo (DeepMind)

    Deep networks plus tree search beat Lee Sedol 4–1 at Go.

    Image: Google DeepMind · Public domain · Wikimedia Commons
  5. 2017
    AlphaZero

    AlphaZero

    Learns chess, shogi and Go from self-play alone, no human games.

    Image: Oliver Orschiedt · CC BY-SA 3.0 · Wikimedia Commons
  6. 2017
    Libratus (Carnegie Mellon)

    Libratus (Carnegie Mellon)

    Beats four top professionals at heads-up no-limit Texas hold'em over 120,000 hands, mastering hidden cards and bluffing.

    Image: Pittsburgh Supercomputing Center · Unclear; (c) PSC, no licence stated · PSC website (psc.edu), Wayback capture 2017
  7. 2019
    MuZero

    MuZero

    Masters games without being told the rules, by learning its own model of them.

    Image: Dietmar Rabich · CC BY-SA 4.0 · Wikimedia Commons

Contested: Torres Quevedo's 1912 El Ajedrecista automaton played chess endgames; Christopher Strachey's 1951 draughts program predates Samuel's.

See it on the timeline