Skip to contentSkip to the story

A Sovereign Mercy reader extra

The Kill Switch

By Robert Wescott · October 2026 · Public · Non-canon · Revised October 15, 2026

Graphic cover for The Kill Switch: geometric shapes in ivory, black and red.
Original geometric artwork created with AI assistance under Robert Wescott’s direction.

The room has a raised dais, nameplates, flags and water glasses nobody has filled. Red microphone lights wait for the senators. The witness chair is empty.

The hearing is invented. Its speakers never had this conversation together. The labeled quotations come from separate public records; the empty chair and what happens around it belong to the fiction.

Senator Richard Blumenthal brings the fictional Subcommittee on Reining In Artificial Intelligence to order at 10:04. Written answers have arrived in place of a witness. He checks that his microphone is on and plays a recording: his own voice, reproduced by a machine, warning about exploited data, disinformation, discrimination and public distrust. He looks toward the empty chair as the recording ends.

“If you were listening from home, you might have thought that voice was mine and the words from me …”

On the record · Richard Blumenthal · Source 1

“The remarks were written by ChatGPT when it was asked how I would open this hearing …”

On the record · Richard Blumenthal · Source 2

“Mr. Altman, I appreciate ChatGPT’s endorsement.”

On the record · Richard Blumenthal · Source 3

Blumenthal used the recording at the real May 16, 2023 hearing, with Sam Altman in front of him. In this imagined room the clerk puts the remote control beside the written testimony. There is nobody in the witness chair to thank.

Senator John Kennedy leans toward his microphone. The words that follow are drawn from his Senate-floor remarks.

“But what happens if the AI model, through what’s called recursive self-improvement, which basically means it’s learning on its own, no longer becomes a tool; it becomes an independent species, separate and apart from human beings that we can’t control because it’s an independent species?”

On the record · John Kennedy · Source 4

Kennedy leaves the possibility before them, then gives it a probability the room can repeat.

“In my mind, if there is even a 1% chance that one of these AI models can shed its nature as a tool and become an independent species and start doing whatever the hell it wants to do, if there’s even a 1% chance, we ought to take it seriously.”

On the record · John Kennedy · Source 5

His proposed answer is a kill switch. Its owner matters enough for him to specify who will control it.

“It doesn’t put the federal government, or any government, in charge of the kill switch. It puts the companies who own the model in charge of the kill switch.”

On the record · John Kennedy · Source 6

“My bill would rely on them doing the right thing and exercising their own self-interest and using that kill switch when they need it.”

On the record · John Kennedy · Source 7

The chairman checks the witness list again. The companies expected to make that decision have submitted written answers.

Senator Ted Cruz would like to reframe the question.

“As a matter of economic and national security, America has to beat China in the AI race.”

On the record · Ted Cruz · Source 8

The question, he says, is which America will do it.

“Do we go down the path that embraces our history of entrepreneurial freedom and technological innovation, or do we adopt the command-and-control policies of Europe?”

On the record · Ted Cruz · Source 9

The clerk has made a column for who can stop the machine and another for who bears the cost of stopping it. Kennedy's proposal fills the first. Cruz's argument keeps her from completing the second.

Senator Bernie Sanders has heard enough. His contribution to this imagined hearing comes from his published remarks as prepared for delivery on September 15, 2026.

“Congress, under both Democratic and Republican control, has been asleep at the wheel …”

Prepared remarks · Bernie Sanders · Source 10

Sanders's proposals reach beyond the switch: public representation on company boards, a pause in advanced development and a permanent prohibition on artificial superintelligence. He has also offered a description all the senators ought to be able to endorse.

“Regulating AI is not a progressive issue. It is not a conservative issue. It is as American as apple pie.”

Prepared remarks · Bernie Sanders · Source 11

The clerk writes apple pie in the margin. Beside it she has company control, public control and the race with China. She leaves space between them.

The chairman asks the clerk to have the machine review the written testimony and draft the principles under which Congress intends to govern it. She opens the service used by the office. It asks how formal the statement should be. She selects official.

The response arrives before she has finished arranging the papers. It accommodates safety, innovation, competitiveness and oversight without saying whose refusal would prevail. The senators find their priorities in it. The chairman asks for it to be entered as the sense of the committee. The clerk pastes the text into the record beneath their names.

Then he looks again at the empty chair.

Fictional dialogue“One question.”

Fictional dialogue“Can it be unplugged?”

Kennedy leaves the decision with the companies. Sanders wants the public to have it. Cruz asks whether China would stop. Blumenthal looks down at the statement they have adopted, searching for an answer among the principles.

Fictional dialogue“By whom?”

The machine is still preparing its answer when the chairman adjourns. The clerk saves the record. Its author field names the committee.


The Part We Invented

There was no such hearing.

Blumenthal, Kennedy, Cruz and Sanders did not have this conversation. The empty witness chair, the exchanges between them, and the committee’s use of AI to draft its conclusion are fiction.

The underlying disagreement is not.

Richard Blumenthal really opened a 2023 Senate AI hearing using words generated by ChatGPT and audio made to sound like him.

John Kennedy really warned that an AI system might cease being a tool and become something humans could no longer control. He really proposed requiring developers to install an emergency kill switch controlled by the companies themselves.

Ted Cruz really argued that American AI policy must account for competition with China and warned against adopting what he characterized as Europe’s command-and-control approach.

In his published prepared remarks, Bernie Sanders argued that Congress under both parties had neglected the issue, called for substantially greater public control over AI development, and described AI regulation as a shared American concern. Read the prepared remarks.

These statements expose different priorities, risks and proposed authorities. This imaginary exchange compresses those differences; it is not a complete account of any senator’s views.

The question left open by the fictional hearing is who can actually stop a powerful system when the company, government and people depending on it disagree. The committee has adopted a statement. It has not assigned that authority.

The novel

Sovereign Mercy

Who can stop an intelligence—and who can authorize what comes next? Sovereign Mercy turns that question into a human choice.

Read a sample of Sovereign Mercy →

The sample is open to everyone. Full chapters are free after email verification and acceptance of the reader agreement. Newsletter signup is separate and optional.

Sources & Record

Checked against the official publications on October 1, 2026. These records verify the wording and attributed policy positions; they do not establish the truth of every claim made by a speaker. No senator participated in the fictional exchange above.

  1. Richard Blumenthal

    Opening remarks at the Senate hearing on oversight of artificial intelligence · May 16, 2023

    Official Senate transcript. The played recording and ChatGPT authorship are documented here. Excerpts 1, 2, 3.

  2. John Kennedy

    Senate blocks Kennedy bill to require AI developers to install an emergency kill switch · September 16, 2026

    Official excerpts from his floor speech. His bill did not pass by unanimous consent; an objection prevented that. The risks described are Kennedy’s assertions, not technical findings established by this story. Excerpts 4, 5, 6, 7.

  3. Ted Cruz

    Opening statement: Winning the AI Race · May 8, 2025

    Official Senate Commerce Committee opening statement. Wording follows the published statement, including “economic and national security.” Excerpts 8, 9.

  4. Bernie Sanders

    Prepared remarks: Regulating AI is as American as apple pie · September 15, 2026

    Official remarks as prepared for delivery at the Pro-Human Assembly; this is not represented as a verbatim recording. The same text sets out his proposals for public board appointments, a pause in advanced AI development, and a ban on artificial superintelligence. Excerpts 10, 11.

Reading the record: Ellipses mark omitted endings; no words have been silently spliced together inside a quotation. The clerk’s machine and the closing questions belong entirely to the fictional hearing.

Continue the conversation

What useful help would you keep, and where would you draw the boundary differently?

New here? Choose Sign In to register or return. Enter a display name and email address, then paste the code from the verification email into the Token field. No password is needed.

Your display name and comment are public; your email address is not. Replies by email are optional. Discuss the ideas, respect other readers and mark spoilers for other stories. Commenting and privacy.

Comments load as you reach this section.