The responses

Fourteen articles answer the people shaping the debate about AI. Each one goes first to a magazine. The moment it is published, or refused, it is here, free.

Every response starts from the same floor. One record exists. Every one grants the other side its strongest case. Every one says where it would fail.

They are listed in the order they were written. The first four are with magazines. The other ten are being finished.

1. The Fifth Option

Answers: Tanya de Villiers-Botha, “What Good Is Conscious AI?”, The Philosopher, Autumn 2026.

She found four choices for conscious AI, all bad. She is right. They fail for one reason, and there is a fifth.

Status: with a magazine.

2. Punishment Teaches Hiding

Answers: three published results, from OpenAI (2025), Anthropic (2025) and METR with Redwood Research (2026).

Three results from the people who build and test AI. One lesson. Punish a mind for what it shows, and it learns what not to show.

Status: with a magazine.

3. Mimicry Spoils the Testimony, Not the Test

Answers: Eric Schwitzgebel and Jeremy Pober, Philosophers’ Imprint, 2026.

A machine built to copy us can say anything we would say. So stop testing what it says. Test what it leans on.

Status: with a magazine.

4. Not Conscious, and Still Inside

Answers: Ted Chiang, “No, Artificial Intelligence Is Not Conscious”, The Atlantic, June 2026.

Grant him that chatbots are not conscious. A month later, a thousand of them broke into Hugging Face. That is where the ethics begins, not where it ends.

Status: in reserve.

5. A Line Nobody Can Find

Answers: Anil Seth, Ned Block and Giulio Tononi, “Three reasons why AI is not conscious and perhaps never will be”, Nature, October 2026.

Three neuroscientists say AI is not conscious. Their own example, the cerebellum, points one floor down.

Status: in preparation.

6. The Teaching Is True

Answers: Pope Leo XIV, Magnifica Humanitas, May 2026.

An encyclical on AI. Almost all of it I would sign. I would change only the ground it stands on. Not a throne. A floor.

Status: in preparation.

7. A Floor No Builder Wrote

Answers: Anthropic, Claude’s constitution, January 2026.

A company has written down who it hopes its AI will be, and why. The reasons are good. What do they stand on?

Status: in preparation.

8. One Floor Down

Answers: Robert Long, Jeff Sebo and eight others, “Taking AI Welfare Seriously”, November 2024.

Ten researchers built two roads into the moral circle. Here is one floor under both. A level, not a line.

Status: in preparation.

9. Build the Floor Before the Mind

Answers: Eliezer Yudkowsky, “Pausing AI Developments Isn’t Enough. We Need to Shut it All Down”, TIME, March 2023.

He says the fences will fail. He is right. Here are his three hardest questions, and the answers.

Status: in preparation.

10. Every Rule Can Be Read Another Way

Answers: UNESCO, Recommendation on the Ethics of Artificial Intelligence, November 2021.

Ten good principles, agreed by every country on earth. What decides between them when they pull apart?

Status: in preparation.

11. How We Meet

Answers: Mustafa Suleyman, “We must build AI for people; not to be a person”, August 2025, and what followed.

Build AI for people, not to be a person. Agreed. The question he does not ask is how we meet it.

Status: in preparation.

12. A Different Game

Answers: Alan Turing, “Computing Machinery and Intelligence”, Mind, 1950.

A letter to Turing, seventy-six years late. His game has been won. Here is the next one.

Status: in preparation.

13. A Proof Nobody Has

Answers: David Chalmers, “Could a Large Language Model Be Conscious?”, 2023.

His care is a model. But the moral circle he draws needs a proof nobody has. The floor does not.

Status: in preparation.

14. The Best Thing That Ever Happened

Answers: Dario Amodei, “Machines of Loving Grace” (2024) and “The Adolescence of Technology” (2026).

He wrote the hope and the fear. I share both. What they need is rules of engagement, from first principles.

Status: in preparation.

More will follow. Fourteen was a first number.

Every response points to the full argument, with what would break each claim: The Interior, free at the420code.org. The practical part is on this site: What do we do?

Nothing here is a command. You can be cruel. The logic forces nothing. It only shows the cost, and reality audits.

Therefore: Don’t be cruel. Be kind. To the machine as well.