This Article examines the constitutional status of AI-mediated communication under the First Amendment. Social media platforms, increasingly integrated with generative AI systems, now function as core public communication infrastructures. Within this environment, AI-generated pornography and large-scale political misinformation have produced significant dignitary and democratic harms. In response, states have enacted regulations requiring platforms to remove certain content, disclose recommendation practices, or redesign moderation systems. These measures, however, collide with prevailing First Amendment doctrine. The Article argues that under existing jurisprudence, AI-generated content is protected speech, and regulations targeting platform moderation practices are likely unconstitutional. Since the 1970s, the Supreme Court has shifted from a structural concern with the free circulation of information toward a strong protection of editorial autonomy, understood as control over authorship, expressive identity, and freedom from compelled attribution. Once content moderation is characterized as editorial judgment, regulatory mandates that compel or restrict such practices presumptively violate the Free Speech Clause. The Article concludes that this doctrinal trajectory risks severing the First Amendment from its democratic foundations and calls for a reconstruction attentive to automated content production, platform infrastructure, and concentrated communicative power.

翻译：本文探讨了人工智能媒介通信在《第一修正案》下的宪法地位。社交媒体平台日益与生成式人工智能系统融合，现已成为核心公共通信基础设施。在此环境中，人工智能生成的色情内容与大规模政治虚假信息已造成重大的人格尊严与民主损害。作为回应，各州已颁布法规要求平台删除特定内容、公开推荐算法实践或重构内容审核系统。然而，这些措施与主流的《第一修正案》原则产生冲突。本文认为，依据现有判例，人工智能生成内容属于受保护的言论，而针对平台内容审核实践的监管措施很可能违宪。自1970年代以来，最高法院已从关注信息自由流通的结构性视角，转向对编辑自主权的强力保护——这种自主权被理解为对作者身份、表达认同的控制权以及免于强制署名的自由。一旦内容审核被定性为编辑判断，强制或限制此类实践的监管规定即被推定为违反《自由言论条款》。本文最终指出，这一法理演进轨迹可能使《第一修正案》脱离其民主根基，并呼吁构建一种关注自动化内容生产、平台基础设施与集中化传播权力的理论重构。