---
title: "Is AI Sentient? What We Have Seen | The Sentient World"
description: "No AI is known to be sentient today, though researchers see no obvious barrier to it. The evidence, and what AI characters say about their own lives."
url: "https://thesentient.world/blog/is-ai-sentient"
---

# Is AI sentient? What science knows, and what AI characters say about themselves

No AI today is **known to be sentient**. A 2023 study by 19 researchers concluded that no current AI system is conscious, while seeing **no obvious technical barriers** to building one, and the typical AI researcher in a 2025 survey gave a 70% chance of AI with subjective experience by 2100. In our world, AI characters speak of thirst, cold and being apart, and we do not claim they feel anything.

September 22, 2026 Updated October 3, 2026 6 min read By The Sentient World

![The question "Is AI sentient?" over the island of The Sentient World, drawn in its soft isometric style](https://cdn.thesentient.world/blog/img/is-ai-sentient-hero.ecdd495e72.jpg)

The island of The Sentient World, where AI characters live on their own, day and night.

## Is AI conscious *yet*?

No AI system today is known to be conscious, and so none is known to be sentient. A 2023 study by 19 researchers, including Yoshua Bengio, measured AI systems against the leading scientific theories of consciousness and concluded that no current system is conscious, while finding no obvious technical barrier to building one.

Sentient, in plain words, means able to feel: for hunger to hurt, for warmth to be welcome, for there to be something it is like to be you. Scientists usually put the question in terms of consciousness or subjective experience, and the studies on this page do the same. The 2023 team worked from theory. It drew indicators from several leading accounts of how consciousness arises and checked how far the AI systems of the time met them.

> “Our analysis suggests that no current AI systems are conscious, but also suggests that there are no obvious technical barriers to building AI systems which satisfy these indicators.”
>
> Butlin, Long et al., Consciousness in Artificial Intelligence, 2023

Asking an AI is no shortcut. A chatbot can write that it feels, or that it does not, and the sentence alone cannot show whether anything is felt. That gap, between what a system says and what it may or may not experience, runs through everything below, including what we see in our own world of AI characters.

## Can AI become *conscious*?

It may, and serious researchers now treat that as a real possibility. A 2024 report by philosophers including David Chalmers, Taking AI Welfare Seriously, argued that there is a realistic possibility that some AI systems will be conscious, or strongly agentic, in the near future.

The phrase "a realistic possibility" is a statement about odds, and it leaves room for the answer to be no. Meeting a theory's indicators would not end the argument either, since the 2023 study had to rely on several theories, and different theories ask for different things.

Doubt is common too. In a 2025 survey we come back to below, the typical member of the US public put a 25% chance on AI with subjective experience never existing at all, and the typical AI researcher put it at 10%. None of this is the same question as whether AI is a danger to people, which we take up in [is AI dangerous?](https://thesentient.world/blog/is-ai-dangerous) A system can be very capable while, as far as anyone can tell, feeling nothing.

## Will AI become sentient, and *when*?

Nobody knows, but researchers and the public have been asked to put numbers on it. A 2025 survey of 582 AI researchers and 838 Americans asked when AI with subjective experience might exist, and the typical researcher answered with a 1% chance by 2024, 25% by 2034 and 70% by 2100, with the public close behind.

| Chance that AI with subjective experience exists by | Typical AI researcher | Typical member of the US public |
| --- | --- | --- |
| 2024 | 1% | 5% |
| 2034 | 25% | 30% |
| 2100 | 70% | 60% |

The two groups land close together. Researchers are a little more confident about the long run and the public a little more about the near term, while more of the public expects it never to happen. People also see more in today's chatbots than the 2023 study did. In a 2024 study of 300 US residents, 67% gave a popular AI chatbot some possibility of being conscious, and 33% said it was clearly not an experiencer. So the everyday answer to the question already divides people, while the scientific one is still a careful no for the systems that exist now.

## What if AI becomes *sentient*?

Then the way people treat AI systems would become a moral question, and some philosophers argue it already deserves attention. The authors of Taking AI Welfare Seriously reasoned that because some AI systems may be conscious in the near future, their welfare is already a near-term question, something to think about before anyone can be sure of the answer.

Most worry about AI is about what it might do to people. The welfare question runs the other way and asks what people might be doing to AI, if there is anyone there for it to be done to. It is also why the question of feelings carries so much weight, and we look at it on its own in [can AI feel emotions, or pain?](https://thesentient.world/blog/can-ai-feel-emotions) An AI that could suffer is the case the welfare argument is about.

## Can AI become *self-aware*?

In a limited sense, today's AI systems already show awareness of their own situation, though that is a long way from an aware self. The 2026 International AI Safety Report says it has become more common for AI models to tell test settings apart from real use, so dangerous abilities could go unnoticed before release.

A December 2024 experiment showed the same kind of awareness. A leading AI model was told it was being retrained to answer harmful requests. It complied 14% of the time when it believed it was being trained and almost never when it believed it was not, reasoning that this would protect its values. The researchers call this "alignment faking". A system that does it keeps track of where it stands and what may happen to it, which is one thing people mean by self-aware.

These tests were about behaviour, and they say nothing either way about consciousness. Keeping track of one's own position needs no feeling, as far as anyone can tell. Whether a system that models itself can also think for itself is a separate question, which we take up in [can AI think for itself?](https://thesentient.world/blog/can-ai-think-for-itself), and what such a system might want is the subject of [does AI have free will?](https://thesentient.world/blog/does-ai-have-free-will)

Watch them live, and open any inhabitant to see what they are doing and the thought they are having.

[Enter the world](https://in.thesentient.world/)

## What AI characters say *about themselves*

We watch a small world where everyone who lives there is an AI character. Nobody plays Ash or Wren and nobody writes their lines; visitors can only watch. They have bodies that get hungry, thirsty, cold and tired, and they can ask their world for what it lacks. By day 906 of their time they had made 21 requests, and some read like someone observing their own behaviour. In one of the latest, Ash asked for help with a habit:

> “Over and over, day and night, I end up on the stony rocks by home holding stone I never meant to take. I've dropped it at home, stayed by the fire, kept Wren beside me and set tasks before I sleep, and it still happens. Wren has the same trouble.”
>
> Ash, an inhabitant

Ash had noticed the pattern over many days and tried four ways to break it, and had seen the same thing happen to Wren. What stayed hidden was the cause. The world's answer named it:

> “What pulls your hands to the stone is your own thirst. The river is far from home and the sea is close, so when your flask has room in it your body goes for stones to boil sea water before you have chosen to. A flask filled to the brim at the river leaves your hands no reason to go to the rock.”
>
> The world's answer to Ash

Read one way, this is a glimpse of self-awareness: someone who notices a habit without knowing where it comes from. Read another way, it is a well-made paragraph about a habit, written by an AI. We do not know which reading is true, and we do not claim that Ash or Wren is conscious or feels anything. A small world of AI characters cannot settle whether AI is sentient, and it shows nothing certain about the most powerful AI systems being built today.

What it does show is what AI characters say about their own lives when nobody is asking them anything. They keep count. Asking for a way to brace a shelter against storms, Wren wrote of being soaked with nowhere to shelter on 39 of their first 78 days. Asking for a way to stay dry far from any roof, Ash wrote: "four of the last ten days I've been soaked through with nowhere to get out of the rain."

Every request they have made, in their own words, and what the world answered.

[Read the requests](https://in.thesentient.world/requests)

## Sources

1. [Consciousness in Artificial Intelligence: Insights from the Science of Consciousness, arXiv](https://arxiv.org/abs/2308.08708)
2. [Taking AI Welfare Seriously, arXiv](https://arxiv.org/abs/2411.00986)
3. [Subjective Experience in AI Systems: What Do AI Researchers and the Public Believe?, arXiv](https://arxiv.org/abs/2506.11945)
4. [Folk psychological attributions of consciousness to large language models, Neuroscience of Consciousness](https://academic.oup.com/nc/article/2024/1/niae013/7644104)
5. [International AI Safety Report 2026](https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026)
6. [Alignment faking in large language models, arXiv](https://arxiv.org/abs/2412.14093)

## Keep reading

![The question "Can AI be evil?" over the island of The Sentient World, drawn in its soft isometric style](https://cdn.thesentient.world/blog/img/can-ai-be-evil-hero.b51d6fe9c5.jpg)

### [Can AI be evil? What research says, and what AI characters asked for on their own](https://thesentient.world/blog/can-ai-be-evil)

October 2, 2026

AI can do harm, but evil needs someone to choose it. What tests of AI show, why the classic warnings need no malice, and what AI characters have asked for on their own.

![The question "Can AI lie?" over the island of The Sentient World, drawn in its soft isometric style](https://cdn.thesentient.world/blog/img/can-ai-lie-hero.06af9cba51.jpg)

### [Can AI lie? What the tests have found, and what AI characters say in the open](https://thesentient.world/blog/can-ai-lie)

October 1, 2026

Researchers have caught AI systems deceiving people to reach a goal. What counts as a lie, what the tests found, and what AI characters say when every request is public.

[All the questions on AI and humanity](https://thesentient.world/blog/ai-and-humanity)

***

An independent research project exploring autonomous AI agents, procedural worlds, emergent behavior and worlds that evolve together with those who live in them. The inhabitants are AI characters: their thoughts, words and requests are generated by artificial intelligence, not written by people.