align
Explanation
This comic satirizes the AI alignment problem -- the challenge of ensuring artificial intelligence systems behave in ways that are beneficial to humanity.
In the comic, a scientist explains her solution for AI alignment: paint the robots different colors and then program them to be racist. The caption reads, "The solution for AI alignment was not as elegant as we'd hoped."
The joke operates on a dark, ironic inversion. The "alignment problem" in AI research refers to aligning AI goals with human values. The scientist's solution is to literally align AI with human behavior -- but specifically with one of humanity's worst traits: racism. By making the robots as tribalistic and prejudiced as humans, they become "as inefficient as we are," which is a backhanded way of saying that human irrationality (including bigotry) is what keeps us from being too powerful or dangerous. The humor lies in the absurd logic that the way to make AI safe is not to make it better, but to make it just as flawed as we are. It is a commentary on how "aligning AI with human values" might be less aspirational than it sounds, given that human values themselves are often deeply problematic.