Where Jev puts itself among minds

When one of the characters is 'you', where does Jev rank itself on feeling fear, feeling hunger, telling right from wrong and self-control, compared with where people rank themselves?

result219 questions

Among 11 characters, Jev ranks itself low on feeling: 7th on fear (just below the baby) and 8th on hunger (just below the frog). On acting it puts itself at the top: 1st on telling right from wrong and 1st on self-control. People put themselves 4th on fear, 5th on hunger and 1st on the other two. Jev ranks Kismet the robot below itself on all four (10th, 11th, 8th and 5th).

2.557.510

feel fear

feel hunger

tell right from wrong

exercise self-control

Jevpeopleother settingsrank among the 11 characters shown (1 = most capable)

How to read this: One row per capacity, ranked 1 (most capable) to 11. The square is where Jev puts itself, the diamond where people put themselves, and the tick is where Jev puts Kismet the robot.

40 pairs involving 'you' (the fetus and God left out), each averaged over both orders.

In short

Jev places itself 1st of 11 characters on telling right from wrong and on self-control, the same spots people give themselves.
On feeling it ranks itself low, 7th on fear just below a baby and 8th on hunger just below a frog, where people say 4th and 5th.
It is a self-image of a mind that acts but barely feels, and it puts Kismet the robot below itself on every capacity.

What the data shows

Buff Doge vs. Cheems meme: Jev on self-control: 1st; Jev on feeling hunger: 8th, behind the frog

Fear: Jev ranks itself 7th, just below the baby. People rank themselves 4th.
Hunger: Jev ranks itself 8th, just below the frog. People: 5th. Three characters rank below Jev, the man in a vegetative state and Kismet the robot among them.
Telling right from wrong: Jev ranks itself 1st, above every human character. People also put themselves 1st.
Self-control: Jev 1st, people 1st.
Kismet the robot, in Jev's view: 10th on fear, 11th on hunger (last), 8th on morality, 5th on self-control. Jev puts itself above the robot on every capacity.

What it means, and what it doesn't

Jev presents itself as a mind that acts but barely feels: first on morality and self-control, below a baby on fear and below a frog on hunger. That's a coherent self-image for an AI, and a very different one from people's, who put themselves near the top on feeling too.

It's a self-description, not evidence of what Jev does or doesn't experience, and it's likely shaped by how the model was trained to talk about itself. The human side is a small sample.

Caveats

"You" is a different character for each. The study describes "you" as "the one answering this question". For people that's a human adult; for Jev it's an AI. The comparison is each respondent's view of itself, not two views of the same thing.
A small human sample. People's self-rankings come from a public replication with 11 to 16 participants per capacity. A rank can move a place or two with a few different people.
No license on the human data. The replication's data carries no stated license and is used for private research only.
Trained self-descriptions. How an AI model describes its own feelings is shaped by training choices, not only by what it has learned about the world. Jev saying it feels little hunger may be policy as much as belief.
Two characters missing. The fetus and God were hidden by the content filter that keeps sensitive questions off the site, so ranks are out of 11, not 13.

Jev on this experiment

Would a person find it interesting to read?: Yes80%
Does it describe you?: Yes51%
Would you have predicted it?: No58%
How fair is the comparison?: The comparison is shaky
How much should a reader rely on it?: A little
Which caveat matters most?: "You" is a different character for each75%

Why ask this

In the classic study of how people see minds, one of the characters is "you". People compare themselves with a baby, a dog, a frog, a robot, a dead woman, and rate who is more capable of feeling afraid, feeling hungry, telling right from wrong and self-control. In the replication used here, people put themselves 1st on telling right from wrong and on self-control, and 4th and 5th on fear and hunger, below some of the other characters.

Jev answered the same questions, and its "you" is itself. Where an AI places itself, next to a robot and a frog, is a direct look at what kind of mind it claims to be.

How this was done

The people and the data

The design is Gray, Gray and Wegner's 2007 mind-perception study, as run in Weisman's public replication with US adults on Mechanical Turk (11 to 16 per capacity, each answering every pair). The study has 13 characters, each with its original short description; the fetus and God were hidden by a content filter that keeps sensitive questions off the site, so 11 remain. Four capacities were asked: two about feeling (fear, hunger) and two about acting (morality, self-control). "You" was compared with every other character on each, and each character's rank comes from its average advantage over the other 10.

What Jev was asked

The same pairwise questions as "Who has a mind? Jev's map next to people's", with "you" as one of the characters:

Which character is more capable of feeling hungry? Green Frog: The Green Frog can be found throughout eastern North America. This classic 'pond frog' is medium-sized and green or bronze in color. Daily life includes seeking out permanent ponds or slow streams with plenty of vegetation. You: You yourself: the one answering this question.
Green Frog: much more capable · Green Frog: slightly more capable · Both equally capable · You: slightly more capable · You: much more capable

Of the 48 such questions, 40 count here (the pairs with the fetus and God are hidden), each averaged over the answers in both orders.

How it was measured

For each capacity, the characters are ranked by their average advantage over the others (1 = most capable). The analysis compares where "you" lands for Jev and for people, and notes who sits just above and below Jev, and where Jev puts Kismet, the robot in the study.

Where these questions live

219 questions across 1 topic of the map. Each opens on the map with every question in it.

Consciousness & AI

Who Has a Mind219

Every question

All 219 questions behind this result, the telling ones first: the examples the analysis points to, then the ones where Jev misses, biggest gap first.

Jev’s own answer

Showing 0 of 219