The speaker, a cartoonist, delves into the potential of AI in humor creation, highlighting the unique human element in humor that AI struggles to replicate. He discusses various humor theories like superiority, incongruity, and benign violation, and how they apply to cartoons. The speaker humorously critiques AI's current limitations in understanding and creating humor, using examples from his work and experiments with AI-generated captions. He emphasizes that while AI can assist in brainstorming, it lacks the human vulnerability and emotional depth that underpin true humor. The speaker also shares insights from a study comparing AI's ability to match captions to cartoons, noting that while AI is improving, it still trails behind human creativity. He concludes with a humorous take on AI's potential threat, suggesting that AI's understanding of humor is far from achieving a human-like sense of humor.