The Ship of Theseus and AI Clones: A Philosophy Primer

1Plutarch’s ship: the original puzzle

The Greek writer Plutarch, in his Life of Theseus (around the year 100), describes the thirty-oared ship on which the hero Theseus returned to Athens. The Athenians kept it for centuries. As timbers rotted, they replaced them with new ones.

Eventually, Plutarch says, the ship became “a standing illustration for the philosophers in the mooted question of growth” (Perrin’s translation). Some argued it was still the same ship; others said it was not.

Philosophers usually separate two meanings of “the same”. Two new phones from a factory are qualitatively the same: they look and work alike. Your phone today and your phone last year are numerically the same: one object that persisted. The ship puzzle is about numerical identity. Is the repaired ship one continuing thing, or a replacement?

2Hobbes’s twist: two ships

In De Corpore (1655), Thomas Hobbes made the puzzle sharper. Suppose someone kept every old plank as it was removed and later put them together again in the same order. Now there are two ships. The repaired one has continuity of shape and history; the rebuilt one has all the original material. Hobbes noted that if both counted as the original, there would be two ships that were numerically one, which is absurd.

Hobbes’s own answer was that it depends on what we are asking about. Asking whether Socrates is the same man is different from asking whether he has the same body. Our answer depends on the kind of thing we have in mind.

Our bodies face a version of this puzzle. Many cell types are renewed over the years, although a 2005 study using carbon dating found that neurons in part of the adult cerebral cortex are as old as the person. Either way, few people think they become someone else because their cells change.

3From planks to neurons: gradual replacement

Philosophers of mind turned the ship into a thought experiment about brains. Imagine replacing one neuron with a tiny device that does exactly the same job, then another, and another, until the whole brain is artificial.

David Chalmers used this scenario in 1995 to argue that consciousness would probably survive. It seems implausible that experience would quietly fade while the person kept behaving and reporting exactly as before. In a 2010 paper he applied it to identity, imagining 1% of a brain replaced each month for 100 months. He concluded that if we survive ordinary change over many years, we plausibly survive gradual uploading too.

This is the Ship of Theseus with a positive answer: continuity of function, step by step, keeps the person. It remains a thought experiment. No technology can replace a living neuron in this way today.

4Copies and teleporters: Parfit’s challenge

Hobbes’s second ship returns in Derek Parfit’s Reasons and Persons (1984). Parfit imagined a teleporter that scans your body on Earth, destroys it, and builds an exact replica on Mars. The person on Mars remembers your life and feels like you. Is it you?

Then Parfit changed the case. In his “branch-line” version the scanner fails to destroy the original, so there are two people with the same past. They cannot both be you, just as Hobbes’s two ships cannot both be the original.

Parfit’s conclusion was radical. He argued that identity is not what matters. What matters is psychological continuity and connectedness: overlapping memories, intentions and character, caused in the right way. Other philosophers disagree, for example those who hold that we are biological organisms. The debate is still open, and fiction keeps returning to it: Transcendence asks whether an uploaded scientist is still himself, and Altered Carbon imagines minds moved between bodies.

5Four kinds of “digital you” compared

Chalmers lists several forms of uploading, including reconstruction from records such as writings, videos and other people’s memories. The table compares these ideas with AI clones that actually exist today.

6So what is an AI clone, philosophically?

Seen through the ship puzzle, an AI clone is not a repaired ship, a rebuilt ship or a teleported one. No plank of the original is touched. You keep living, changing and making new memories, while the clone stays a separate thing built from records of you.

A better image is a model ship made from photographs and the captain’s logbook. It can be remarkably faithful in style and detail, and it can be precious. But nobody would say the original ship sailed into the model.

This matters for honesty. Calling a clone a “continuation” or “digital immortality” suggests something no current technology delivers. Our article Is an AI clone really you? looks at what a clone captures and lacks in everyday terms, and digital immortality explained sorts the real products from the science fiction.

7Where Memory Clone fits

Memory Clone, a Windows app currently in alpha, builds exactly this kind of modest model: a private AI simulation of how a person writes, learned from chats you export yourself and notes you add. It learns style separately for each relationship and keeps imported originals on your PC. Every clone is presented as an AI simulation, not as the person.

If you would like to try a clone of yourself and see the difference between a portrait and a person, the download page explains how to start. Our guide What is an AI clone? gives a plain definition.