Debate Is For the Silent Audience

Public arguments are not primarily contests between the two visible disputants but performances meant to persuade a third, silent audience who compares competing cases. Large language models can manufacture plausible-sounding positions, but because they lack adversarial testing and social judgment, their arguments risk filling the public sphere with untethered rhetoric that looks persuasive but hasn't survived scrutiny. — If true, this shifts how we should regulate, design, and use AI argument tools: focus less on policing content and more on preserving adversarial testing, provenance, and cues that signal which claims have been meaningfully contested.