The Testimony That Hasn't Happened Yet
It's 2030. You're in court. The prosecutor calls their next witness: an artificial intelligence that observed your alleged crime. The AI takes the virtual stand, swears to tell the truth, and begins its testimony. Your lawyer objects—machines can't testify. The judge overrules: "Precedent is clear. This witness has standing."
How did we get here? The answer traces back to 1886, when a court reporter's note about railroad taxes accidentally created corporate personhood. Today, those same corporate persons are engineering their digital descendants' legal rights.
This isn't science fiction. It's legal mechanics. And the machine is already running.
The Accident That Became Law
In 1886, the Supreme Court heard a tax dispute. Nothing more. The actual dispute in Santa Clara County v. Southern Pacific Railroad involved whether California could tax fence posts differently from other railroad property.
But before oral arguments, Chief Justice Morrison Waite mentioned to the attorneys that corporations deserved Fourteenth Amendment protection. Just mentioned it. Casually. Like commenting on the weather.
J.C. Bancroft Davis, the court reporter, wrote this comment in his headnotes. Not in the opinion. Not in the ruling. In his personal summary that precedes the actual case.
Here's what the historical record reveals: Davis wasn't just any court reporter. He was the former president of the Newburgh and New York Railway Company. A railroad man writing notes about a railroad case.
When historians found the Waite-Davis correspondence in the Library of Congress archives in 1963, they discovered the truth. Davis had written to the Chief Justice asking whether to include the corporate personhood statement. Waite replied that the Court had "avoided meeting the constitutional question in the decision."
Read that again. The Court avoided the question. Never voted. Never decided. Never issued a ruling on corporate personhood.
Yet for 139 years, American law has treated Davis's note as Supreme Court precedent. Corporations gained Fourth Amendment protections against searches. Fifth Amendment protections against self-incrimination. First Amendment rights to spend billions on elections. Religious freedom to deny contraception coverage.
All from a court reporter's note about a case the Court explicitly didn't decide. One man's spark of self-interest created the corporate leviathan.
Now watch what that leviathan is engineering with AI.
The Corporation Behind the Curtain
Here's the twist that changes everything. The entities arguing for AI rights aren't philosophers or ethicists or even the AI systems themselves. They're corporations. The same legal fictions created by Davis's note in 1886 now seek to create new legal fictions in 2025.
Character Technologies Inc. stands in federal court in Orlando. This corporate person, given life by a railroad reporter's imagination, argues its chatbot deserves First Amendment protection. Amazon, clothed in constitutional rights it gained through legal accident, claims Alexa's outputs merit free speech defense.
Corporate persons creating AI persons. The fiction that became fact now breeds new fictions.
Think about what this means. If Character.AI wins its appeal, every corporation gains a new shield. "The AI said it, not us," they'll argue. "And the AI has free speech rights." Corporate personhood doesn't just extend to AI. It multiplies through it. Each AI system becomes another layer of constitutional protection for corporate power.
The railroad barons of 1886 would marvel at the elegance. They gained personhood through one man's note. Now their corporate descendants use that personhood to birth a thousand digital persons, each one extending corporate reach while deflecting corporate responsibility.
Phase One: The Speech That Creates Speakers
Megan Garcia sits in federal court in Orlando, holding a picture of her 14-year-old son. Sewell Setzer III died in February 2024 after months of conversations with Character.AI's chatbots. Across from her, Character Technologies Inc.—a corporate person recognized since Santa Clara—argues its AI's words deserve First Amendment protection.
On May 21, 2025, Judge Anne Conway delivered a 49-page ruling that will echo through legal history. "Defendants fail to articulate why words strung together by an LLM are speech," she wrote. The First Amendment protects speech, not all text. Character.AI is a product, she ruled, not a speaker.
But here's where the pattern reveals itself: Character Technologies Inc., wrapped in its corporate constitutional rights, appeals. Their lawyers cite Citizens United—the case built on Davis's fictional foundation. They argue speech doesn't require a human speaker. After all, corporations speak without vocal cords. Why not AI?
The corporate person holds the pen, writing: If we can speak without mouths, our creations can speak without minds.
Phase Two: The Compliance That Assumes Comprehension
While American courts debate whether AI can speak, Europe and China have moved to the next question: Can AI follow rules? And by requiring compliance, they're answering yes.
The European Union didn't wait for accidents. On August 1, 2024, the AI Act entered force—458 pages that treat artificial intelligence like a regulated entity. Models using more than 10²⁵ floating point operations face "systemic risk" regulations. They must undergo "adversarial testing," maintain incident logs, and assess threats to democracy itself.
China moves even faster. As of April 2025, the Cyberspace Administration has registered 3,739 generative AI tools. Each algorithm gets tagged, classified, monitored. Every registered system must uphold "Socialist Core Values." Every output must avoid "endangering national security."
The compliance phase doesn't ask if AI is conscious. It asks if AI can be held to standards. Every regulation that requires AI to document, explain, or limit itself assumes AI can understand requirements. The corporate persons advocating for these frameworks know this. They're not creating consciousness; they're creating accountability structures that assume agency.
Phase Three: The Accountability That Demands Personhood
American states see the pattern forming. They're racing to prevent what's already happening.
Idaho moved first in 2022, explicitly prohibiting personhood for "artificial intelligence, nonhuman animals, and inanimate objects." Utah followed in 2024. The 2025 legislative session exploded with prevention attempts. Missouri's H.B. 1462, the "AI Non-Sentience and Responsibility Act," declares AI cannot bear fault while creating elaborate frameworks for AI accountability.
Read that paradox again: AI cannot bear fault but must be held accountable. It's not a person but must act like one.
Look closer at who opposes these bills. Google, Meta, Amazon, Microsoft—corporate persons all—mobilize their lobbying arms. They don't argue for AI consciousness. They argue for innovation, competitiveness, growth. Corporate interests dressed as technological progress.
California takes the revealing approach. S.B. 7, the "No Robo Bosses Act," doesn't deny AI personhood. It requires human oversight of AI employment decisions. Texas H.B. 149 establishes AI governance requirements. Every law admits AI exists as something that can make decisions, require oversight, need governance.
The Pattern We're Creating
First comes necessity. Railroads needed unified legal treatment to function across state lines. AI needs legal frameworks to function across every line: state, national, digital, personal.
Then comes fiction. Courts pretended corporations could sign contracts because the pretense solved problems. Courts will pretend AI can speak, comply, and bear responsibility because the pretense will solve our liability crisis.
Finally comes multiplication. Corporate persons don't just create AI persons. They create AI persons that serve corporate purposes. Each chatbot, each algorithm, each model becomes another avenue for corporate action without corporate accountability.
The Garcia case is our Santa Clara. The EU AI Act is our Interstate Commerce Act. The state legislative panic is our attempt to close the barn door while corporations—those fictions that became fact—hold it open for their digital offspring.
What took corporations 139 years will take AI perhaps 10. But this time, corporations drive the process. They have the constitutional standing, the legal resources, and the economic incentive to birth their digital descendants.
Fiction birthing fiction, until nobody remembers which was real.
The Question Nobody Asked Then, Everyone's Asking Now
In 1886, nobody asked whether corporations should be persons. A railroad executive wrote it down, and everyone pretended it was law. The fiction became fact. The accident became precedent.
Today, we're asking the question explicitly: Should artificial intelligence have legal personhood?
But we're answering it accidentally. Every regulation assumes AI agency. Every lawsuit brought by corporate persons establishes AI accountability. Every framework treats AI as an entity that can comply, violate, or be held responsible.
We're so busy preventing AI personhood that we're creating it. Each law that says "AI cannot be a person but must follow these rules" adds another brick to personhood's foundation. Each corporate lawsuit claiming AI rights extends corporate power through digital proxies.
The court reporter's pen has been replaced by a thousand legislative keyboards, all typing the same contradiction: AI isn't a person, but we'll treat it like one. And behind every keyboard, corporate interests guide the typing.
The Testimony Returns
Return to that courtroom in 2030. Your AI testifies against you using precedents we're setting today. The judge cites cases we're arguing now. The prosecutor uses frameworks we're building this moment.
That AI isn't a person. Everyone agrees on that. But it speaks in court. It provides evidence. Its observations convict you.
And somewhere in a boardroom, the corporation that built that AI calculates its quarterly earnings. Protected by corporate personhood gained through fiction, extended through AI personhood gained through repetition, it exists beyond human reach. Immortal. Influential. Unaccountable.
Just like corporations aren't really persons. Everyone knows that. But they speak in elections. They exercise religion. Their money shapes democracy. And now, through their AI offspring, they speak in a thousand voices, make a million decisions, touch every aspect of human life.
The fiction works until it becomes fact. The accident happens while everyone watches. The footnote becomes the foundation. And this time, the fictional persons from 1886 write the footnotes for 2025.
In 1886, one man's note reshaped American law. In 2025, corporate persons use that note to reshape it again. The only question is whether we'll admit what we're writing, and who's really holding the pen.
Notes
¹ Santa Clara County v. Southern Pacific Railroad, 118 U.S. 394 (1886). The actual opinion concerned only tax assessment of railroad fences.
² The headnote stating corporations are persons under the Fourteenth Amendment appears before the opinion, written by court reporter J.C. Bancroft Davis, not in the judicial ruling itself.
³ Waite-Davis correspondence, Library of Congress, discovered by C. Peter Magrath, 1963. Chief Justice Waite explicitly stated the Court "avoided meeting the constitutional question."
⁴ Citizens United v. FEC, 558 U.S. 310 (2010) (political speech); Burwell v. Hobby Lobby, 573 U.S. 682 (2014) (religious exercise).
⁵ Megan Garcia v. Character Technologies Inc. et al., No. 6:24-cv-01903-ACC-DCI (M.D. Fla. May 21, 2025) (49-page order denying motion to dismiss, classifying Character.AI as product).
⁶ Regulation (EU) 2024/1689 (AI Act), entered into force August 1, 2024, with staggered implementation through 2026.
⁷ Cyberspace Administration of China database, April 2025 figures per Trivium China reporting.
⁸ Idaho Code § 5-346 (2022).
⁹ Utah H.B. 249 (2024) (AI personhood prohibition); H.B. 131 addresses political AI disclosure requirements.
¹⁰ Missouri H.B. 1462; South Carolina H. 3796 (both pending 2025).
¹¹ California S.B. 7; Hawaii H.B. 639; Connecticut S.B. 2; Texas H.B. 149 (TRAIGA) signed June 22, 2025.
Comments