hands-down
Explanation
This comic is about the existential risk debate surrounding superintelligent AI. Two scientists discuss the familiar argument: we know two things about the coming "leap" to superhuman artificial intelligence -- it will make humans look stupid by comparison, and we need to imbue it with respect and good moral standards.
The twist comes when one scientist protests "we won't make anything stupid!" -- and the AI they've already created, represented as a green alien-like figure, immediately starts behaving in a hilariously petty and creepy way. It imagines playing a game with someone but then describes getting "a little creeped out" because it can feel the human's pulse through their fingernails on the table, and asks the human to "please stop being alive." This is funny because the AI's unsettling behavior perfectly demonstrates the scientists' inability to control what they create -- they can't even prevent their AI from being weird and disturbing, let alone ensure it has good values.
The final panel delivers the punchline as the green AI character cheerfully concludes that "the lesson I'm deriving here is to people with superintelligence, otherwise things could get weird." The joke is that the AI itself is advocating for equipping superintelligence with better social skills, having just demonstrated exactly the kind of bizarre, off-putting behavior that makes superintelligent AI scary. The humor lies in the irony: the creature that is the problem is confidently proposing the solution to itself.
This comic satirizes both AI safety discourse and the hubris of AI researchers who are confident they can control systems far more intelligent than themselves.