This article argues that Section 230’s immunity, which has protected tech platforms from liability for third-party content for 30 years, should not extend to unique generative AI outputs. When AI systems create novel statements that never existed before, they act as content creators – not traditional platforms that merely host or moderate user-generated content. This distinction matters because when AI generates defamatory falsehoods, there’s no third-party human speaker for victims to sue, creating an unprecedented accountability gap. I contend that Section 230’s text never promised immunity for a platform’s own original speech, and courts should recognize this critical difference as they confront AI defamation cases.
