go-ai
Explanation
The Joke
A man recounts the history of AI beating humans at increasingly complex games: first checkers, then chess, then Go ("even mates at a species"). He then proposes a solution: design a game specifically so that humans will always have an advantage over computers. In the next panel, the man sits across from a computer and says, "Okay, computer, let's play. What number am I thinking of?" The computer responds with "5" and the man replies, "Nope, it was 1."
The joke is that the "game" the man invented where humans always win is one with no verifiable rules -- essentially, a game where you can just lie. The computer has no way to check what number the man was actually thinking of, so the human can always claim the computer guessed wrong. It is not a real contest of intelligence; it is simply a rigged game where honesty is optional for one side.
The Humor
The comic plays on the anxieties around AI surpassing human intelligence in game after game (checkers in 1994, chess in 1997, Go in 2016 with AlphaGo). The humor comes from the absurd anticlimactic solution: instead of humans getting smarter or building better strategies, the man just invents a game that relies on the honor system and then cheats. It is a satire of how humans respond to being outperformed -- not by improving, but by moving the goalposts or changing the rules entirely.
There is also a layer of commentary about what actually distinguishes human cognition from machine intelligence: here, it is not creativity or intuition, but the capacity for deception and bad faith.
References
The comic references the history of AI game-playing milestones: IBM's Deep Blue defeating chess world champion Garry Kasparov in 1997, and Google DeepMind's AlphaGo defeating Go champion Lee Sedol in March 2016 -- just weeks before this comic was published on April 9, 2016. The AlphaGo victory was a landmark moment because Go was considered far more complex than chess and was thought to be decades away from AI mastery.