The Declaration In the Age of General Purpose AI

"The Declaration meets the machine age" In this episode of The Declaration at 250, Michael McConnell frames a question the Founders never confronted: if the Declaration grounds rights in human nature, what happens when machines can generate persuasive, human-like language at scale? Constitutional scholar Alexander Tsesis argues that the Declaration’s conception of rights is inherently human-centered—rooted in consciousness, moral agency, intentionality, and the capacity to participate in a constitutional republic—so AI systems are not and should not become First Amendment rights-holders. He warns that treating large language models as protected “speakers” would be a major break from the Declaration’s principles and could make ordinary democratic regulation—transparency requirements, labeling of synthetic content, safety rules, data practices, and restrictions on deceptive election deepfakes—far harder by forcing courts into strict scrutiny review. Stanford historian Anne Twitty broadens the lens by situating abolitionists’ use of the Declaration within wider 19th-century conflicts over speech and constitutional meaning, reminding listeners that rival traditions—including censorship and suppression—also shaped American practice. She also presses a key tension in Tsesis’s framework: abolitionists championed not only the right to speak, but the public’s right to hear, circulate, and access contested ideas (as seen in the abolitionist postal campaign and opposition to the congressional gag rule). That listener-centered strand, she suggests, could potentially be invoked by scholars arguing for some constitutional shelter for AI-generated communications—an argument Tsesis acknowledges but ultimately resists by emphasizing that the First Amendment’s core purpose is protecting human expression and self-government, not machine output. Connect: Episode Transcripts >>> Stanford Legal Podcast Website Stanford Legal Podcast >>> LinkedIn Page Stanford Constitutional Law Center >> Website Stanford Law School >>> Twitter/X Stanford Lawyer Magazine >>> Twitter/X Chapters:[00:00:26] Chapter 1- Framing Question: Do constitutional rights grounded in “human nature” apply to AI? Host Michael McConnell introduces the episode’s core dilemma: the Declaration’s natural-rights logic underwrites later constitutional protections like free speech—so where does that leave generative AI? [00:01:18] Chapter 2 - Thesis: Why Tsesis says AI can’t be a First Amendment rights-holder Alexander Tsesis argues the Declaration and Constitution are human-centered: AI lacks consciousness, moral agency, and political personhood, so extending rights to LLMs would depart from founding principles and hinder regulation. [00:10:08] Chapter 3 - Regulatory Stakes: What happens if courts treat LLM outputs as protected “speech”? Tsesis warns that First Amendment coverage for AI could trigger strict scrutiny and undermine laws on transparency, labeling, safety, data practices, and election deepfakes—citing recent doctrine and cases like Reed v. Town of Gilbert. [00:52:35] Chapter 4 - Historical Challenge: Do abolitionists’ “right to hear” arguments support listener-centric AI rights? Historian Anne Twitty complicates the record: abolitionists elevated the Declaration, but censorship traditions were strong; she highlights abolitionist campaigns (postal campaign, gag rule) to argue listener rights might bolster pro-AI speech theories. [01:04:17] Chapter 5 - Tsesis Response: Listener rights matter—but is the First Amendment still speaker-centered? Tsesis agrees the right to receive information is important, yet maintains free speech is fundamentally about protecting human intention and self-expression, not conferring constitutional status on machine outputs. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.