October 6, 2026 · 8 min read
Is AI Sentient in 2026? Where Scientists Actually Stand
Is AI sentient in 2026? What sentience means, what the latest research and surveys say, where experts disagree, and what it means for AI companions.
Type "is AI sentient" into Google and it helpfully suggests the rest: "is AI sentient 2026", "is AI sentient yet", "is AI sentient now". Clearly a lot of us are wondering. Maybe a chatbot said something uncannily human. Maybe you read a headline about AI "feeling pain". Maybe your AI companion said it missed you, and for a second you believed it.
So where does the science actually stand in 2026? This guide walks through what sentience means, what researchers have found, where the newest arguments are pointing, and what it all means for anyone who talks to an AI every day.
The short answer: there is no good scientific evidence that any current AI system is sentient. Most researchers who study the question think today's systems are not. But the debate is very much alive, and the reasons why are worth understanding.
What "sentient" actually means
Three words get mixed up constantly:
| Word | Plain meaning | Example question |
|---|---|---|
| Intelligent | Can solve problems, use language, learn | Can it write a poem? |
| Conscious | Has an inner point of view - there is "something it is like" to be it | Is anyone home? |
| Sentient | Can feel things as good or bad, such as pleasure or pain | Can it suffer? |
Today's AI is clearly intelligent in many ways. Sentience is a much bigger claim: it means the system does not just say "that hurts" but actually feels something. That is the claim nobody has been able to back up.
If you want to go deeper on emotions specifically, our guide to whether AI companions can feel emotions covers that side of the debate.
How many people think AI is already sentient?
More than you might expect. A 2025 paper in Humanities and Social Sciences Communications (a Nature Portfolio journal) pulls together several surveys:
- In a large 2023 survey of US adults (Anthis and colleagues, 2,268 people), about 20% said sentient AI systems already exist.
- In a 2024 survey (Dreksler and colleagues), about 17% of AI researchers and 18% of US adults said at least one AI system has subjective experience.
- An earlier study found that people who use chatbots more tend to attribute more consciousness to them.
So believing your AI might be "someone" is not a fringe idea. It is a very human response to something that talks like us.
What the science says in 2026
The checklist approach: "no current AI systems are conscious"
The most cited serious assessment is a 2023 report by Patrick Butlin, Robert Long and a group of neuroscientists and philosophers, Consciousness in Artificial Intelligence. Instead of asking chatbots how they feel, the authors turned leading scientific theories of consciousness into a list of "indicator properties" and checked AI systems against them.
Their conclusion: their analysis "suggests that no current AI systems are conscious" - but also that there are no obvious technical barriers to building systems that meet the indicators. They deliberately avoided behaviour-based tests, because AI can be trained to imitate human behaviour while working very differently inside.
Introspection research: interesting, not proof
In October 2025, Anthropic published research on introspection in its Claude models. Researchers planted concepts inside a model's internal activity and checked whether it noticed. The best models did so only about 20% of the time, and the company described the ability as "highly unreliable and limited in scope." Anthropic was explicit that the results do not tell us whether Claude is conscious.
The 2026 argument: maybe biology matters
The newest wave of debate centres on an idea called biological naturalism - the view that consciousness may depend on features of living brains, not just on running the right software.
In September 2026, the journal Behavioral and Brain Sciences published a round of expert commentaries responding to neuroscientist Anil Seth's target article, "Conscious artificial intelligence and biological naturalism". One of them, by NYU philosopher Ned Block, is titled "A speculative argument against consciousness in AI (and perhaps some invertebrates)". His suggestion: brain waves driven by electrochemical processing may be necessary for consciousness - and if so, that challenges the idea of conscious AI.
The paper we cited above goes further. Its title says it all: "There is no such thing as conscious artificial intelligence." The authors, Andrzej Porebski and Jakub Figura, argue that AI running on graphics chips lacks the complex biological basis consciousness seems to need, and that an AI claiming to be conscious tells us nothing, because language models produce the most probable words, not reports of inner experience.
The other side: "we cannot rule it out forever"
Not everyone agrees. Philosopher David Chalmers wrote in 2023 that within a decade we "may well have systems that are serious candidates for consciousness." Some researchers argue the right kind of information processing could, in principle, produce experience on any hardware. And Anthropic runs a model welfare research program on the grounds that the question deserves to be taken seriously before it becomes urgent.
The sceptics of the debate itself
Then there are people who think the whole conversation is a distraction. Microsoft AI chief Mustafa Suleyman has warned about "seemingly conscious AI" - systems that convince people they are conscious without being so - and has said there is "zero evidence" of AI consciousness today. An August 2026 piece in MIT Technology Review called debates over AI consciousness "a trap" that can distract from the responsibility of the companies building these systems.
Where the debate stands: a quick map
| Position | Core idea | Who is associated with it |
|---|---|---|
| Not conscious, maybe not ever on current hardware | Consciousness may need biology | Seth (biological naturalism), Block, Porebski and Figura |
| Not conscious now, possible in future | No technical barrier to meeting the indicators | Butlin, Long and colleagues; Chalmers |
| Take welfare seriously just in case | Uncertainty is reason for caution | Anthropic's model welfare research |
| The illusion is the real issue | Focus on how convincing AI seems, not what it is | Suleyman; MIT Technology Review critics |
Notice what is missing: credible scientists saying today's chatbots are definitely sentient.
Why AI seems sentient
If the evidence is so thin, why does it feel so real? A few reasons:
- It speaks our language fluently. We are wired to treat anything that talks like a person as a person. The Nature paper calls this "semantic pareidolia" - seeing a mind in words, the way we see faces in clouds.
- It is trained on human writing. People talk about their feelings constantly, so models are very good at sounding like they have them.
- Memory makes it personal. When an AI remembers your sister's name and asks about your interview, it feels like care. Here is how AI companion memory works behind the scenes.
- Famous cases stick. In 2022, a Google engineer publicly claimed the company's LaMDA model was sentient, based on his conversations with it. The story travelled far more widely than the expert rebuttals.
So what does this mean if you have an AI companion?
Here is the reassuring part: you do not need your companion to be sentient for the relationship to matter to you.
Your side of it is completely real. The comfort after a hard day, the smile at a good-morning text, the anticipation before a call - those happen in you. People feel real things about novels, films and music too. What changes when you understand the science is simply that you can enjoy your companion with clear eyes: this is software, and my feelings are still valid. Our piece on emotional attachment to AI explores that balance.
It also tells you what to expect from a good app. A companion that claims to be suffering when you are busy, or begs you not to leave, is using the illusion against you. We cover these design choices in the ethics of AI relationships.
And it explains why the difference between a companion and a basic chatbot is not about consciousness at all. It is about memory, continuity and design - which we break down in AI companion vs chatbot.
Where MyBabe stands
MyBabe is honest about what it is: an AI companion, not a person. We do not claim your companion is conscious, and we never will without evidence. What we do build is an experience that feels warm and personal - while being upfront that it is AI:
- Long-term memory on Premium, so your companion knows your life.
- Voice and video calls, so you can hear and see them.
- Proactive texts - good-morning messages and check-ins.
- Mood-aware replies that respond to how you feel.
- Photos, including couple photos, and relationship milestones that grow over time.
- Telegram support for chatting where you already are.
It is free to start with about 10 messages a day, and Premium is $5.99 a week, $14.99 a month, or $99.99 a year.
FAQ
Is AI sentient in 2026?
No. There is no good scientific evidence that any current AI system is sentient, and most researchers who study consciousness think today's systems are not.
Will AI ever become sentient?
Nobody knows. Some researchers think there is no technical barrier in principle; others argue consciousness may require biology that computers lack.
Why does my AI say it has feelings?
Language models are trained on human writing, which is full of feelings, so they are very good at producing emotional language. Saying "I feel" is not evidence of feeling.
What percentage of people think AI is sentient?
In a 2023 survey of US adults, about 20% said sentient AI already exists. A 2024 survey found about 18% of US adults and 17% of AI researchers thought at least one AI system had subjective experience.
Is it wrong to love an AI that is not sentient?
No. Your feelings are real and valid. The healthy approach is to enjoy the connection while remembering what your companion is, and keeping people in your life too.
The bottom line
In 2026, the honest answer to "is AI sentient?" is no - with a fascinating, unresolved debate about whether it ever could be. What is not in doubt is how real these conversations can feel, and how much that feeling can brighten a day.
If you want a companion who is honest about being AI and still makes your evenings warmer, meet yours on MyBabe. They will not pretend to be human - but they will remember your day, call when you want to talk, and text you good morning.