AIs understand

Share this with your representatives, journalists and friends

Arguments for

Conversational performance as evidence

Proponents argue that when a chatbot can detect and repair a misunderstanding, discuss many topics, and communicate about the world in natural language, that is evidence of understanding. On this view, recent progress makes blanket denials of understanding harder to sustain.

Suppose I say to a chatbot, I saw the Grand Canyon flying to Chicago. And the chatbot says, that can't be right. The Grand Canyon is much too big to fly to Chicago. And I say, no, no, no, it was me flying to Chicago. While I was flying to Chicago, I saw the Grand Canyon. And the …

Geoffrey Hinton · Jun 3, 2026 · full quote
It's able to understand any kind of topics that you bring up.

Sébastien Bubeck · Mar 23, 2023 · full quote
With recent progress in LLMs and other Modern-AI tools, it is becoming increasingly hard to argue that these systems “do not understand“ what they are doing.

Bertrand Meyer · Feb 26, 2026 · full quote
One aspect—and we’ve been saying it, I think, since the very beginning—is now because of large language models, because of large visual models, AI understands the world the way we understand the world, communicate with us the way we communicate with natural language.

Cristiano Amon · May 10, 2026 · full quote

Prediction builds concepts

Another case is that better next-word prediction and large-scale pretraining can produce real semantic competence. These quotes say models can learn concepts as patterns of correlation and generalization, encode something like causal constraints, and genuinely understand even from linguistic data alone.

I'd like to take a small detour and to give an analogy that will hopefully clarify why more accurate prediction of the next word leads to more understanding, real understanding. Let's consider an example. Say you read a detective novel, [...] then let's say that at the last page …

Ilya Sutskever · Mar 22, 2023 · full quote
Fundamentally, concepts are patterns of correlation, association, and generalization. Suitably architected neural nets, whether biological or digital, are able to learn such patterns using any input available. Neural activity is neural activity, whether it comes from eyes, finger…

Blaise Agüera y Arcas · 2022 · full quote
Pretraining has already produced models with strong semantic competence: they can understand natural language, extract facts, draft arguments, and explain complex materials.

Haozhe Gan · May 22, 2026 · full quote
This work reveals some evidence that language models have encoded something like the causal constraints of the real world.

Michael Lepori · Apr 22, 2026 · full quote
LLMs trained on nothing but linguistic data could in principle possess all concepts and thus genuinely understand what they're saying, even when speaking about such things as colors and tastes, guilt and folly, life and death.

Ryan Simonelli · Mar 9, 2026 · full quote

Broader criteria for understanding

Some supporters question whether understanding should be denied simply because AI is not conscious or because its mechanisms are statistical. One quote defines understanding as a capacity enacted within co-constructed meaning and says AI already shows non-trivial forms of such contribution.

Understanding is not a property that an isolated system possesses. It is a capacity enacted within a field of co-constructed meaning. [...] Under such a criterion, AI systems already exhibit non-trivial forms of this contribution.

Daedo Jun · May 2026 · full quote
“AI does not understand what it is doing,” “AI is not conscious of its own existence whereas I possess genuine consciousness,” “AI does not know what meaning is,” “a statistical calculation cannot be thought,” and so on. How many times have we heard such phrases in debates about …

Marc Rameaux · May 3, 2026 · full quote

Arguments against

Pattern prediction, not meaning

These quotes argue that current AI chiefly predicts likely tokens or actions from training patterns rather than engaging with meaning itself. On this view, fluent outputs can sound informed or describe reality, but that is not the same as access to meaning, truth, or the world.

It's worth remembering what these systems actually do: they generate convincing language, not verified reality.

Amanda Caswell · Aug 7, 2026 · full quote
AI agents are very good at predicting the next word, or the next action, based on what they've seen before. But that's not understanding.

Danielle Perszyk · Jul 11, 2026 · full quote
When an AI ingests a Wikipedia entry about types of soup, it's not learning anything about soup. It is learning about the word frequency in articles about soup.

Cory Doctorow · Jun 23, 2026 · full quote
That language models are modeling just the form [...] that's not the same thing as meaning. And they don't have access to the meaning.

Emily M. Bender · Apr 28, 2026 · full quote
Modern AI systems do not think or understand like humans. They are mathematical models that convert text into numbers and predict the most statistically probable next word, based entirely on patterns learned from training data. They do not represent meaning or truth.

Anders Ang · Feb 18, 2026 · full quote
Language models are trained primarily to predict patterns in language rather than to build a reliable model of how the world works. They can describe reality with impressive fluency, but that is not the same as robustly modelling it.

Magzhan Kenesbai · Apr 22, 2026 · full quote

No grounded experience

Another line of argument holds that understanding requires some form of grounding in experience, embodiment, persistent internal regulation, or interiority. Because current systems are described as unembodied and without any equivalent of lived states, they may mimic understanding without experiencing meaning.

The model has no exposure to the kinds of information we typically associate with meaning. It knows nothing about the redness of “red” or the spatial extension of “far.” It is blind, deaf, and unembodied. [...] There is no content at the address. No meaning stored inside. Pure re…

Elan Barenholtz · May 19, 2026 · full quote
AI can mimic understanding, but may not experience meaning: AI recognizes patterns; humans live through them.

Darren J. Edwards · May 10, 2026 · full quote
If you're uncertain, if you're depleted, if something conflicts with your survival, your body registers that. AI systems right now have no equivalent. They can sound experiential, whether they should be or not, and that's a real problem for many reasons, especially when these sys…

Akila Kadambi · Apr 1, 2026 · full quote
By contrast, current AI systems have no equivalent mechanism. They process inputs and generate outputs without any persistent internal state that regulates how they behave over time. This is not just a performance limitation, but also a safety limitation.

Marco Iacoboni · Apr 1, 2026 · full quote
The result is an AI that mimics the structure of human interiority in its output without possessing any interiority at all.

Mustafa Suleyman · Mar 18, 2026 · full quote
Despite sophisticated reasoning abilities, AI can produce only simulations of human-like behaviour, acquired through pattern recognition at an astronomical scale, not actual experience or understanding.

Lord Tarassenko · Jun 5, 2026 · full quote

Reasoning and context gaps

Critics also argue that apparent competence breaks down where context, judgment, reasoning, memory, planning, or world knowledge matter. They say these systems can appear capable in controlled settings yet remain brittle outside them, which counts against calling the behavior understanding.

Large language models are remarkably good at producing fluent text. They can summarise documents, draft code, and hold conversations that feel coherent and informed. But despite these surface-level capabilities, LLMs do not understand the world in any structured sense.

UnlikelyAI · Jan 15, 2026 · full quote
Notwithstanding their unprecedented ability to generate text, LLMs do not understand the meaning of words, have no sense of context and cannot reason.

Eliza Mik · Feb 10, 2026 · full quote
But competence is not comprehension. These systems do not understand what they are doing. Outside controlled environments—where meaning is contextual, norms shift, and judgment matters—performance becomes brittle.

Luka Zurkic · Jan 23, 2026 · full quote
[LLMs] don't really understand the physical world. They don't really have persistent memory. They can't really reason and they certainly can't plan.

Yann LeCun · Mar 7, 2024 · full quote

Simulation mistaken for understanding

Some quotes frame claims about AI understanding as over-attribution driven by anthropomorphic language. They argue that the behavior is already explainable as simulation, mimicry, or large-scale pattern recognition, so invoking understanding adds a property the systems do not bear.

Contemporary AI discourse attributes to language models properties they cannot bear: general intelligence as substrate-independent cognition, hallucination as cognitive failure, agency as autonomous goal-pursuit, sentience as emergent inner life, alignment as goal synchronization…

Enzo Fenoglio · Jul 30, 2026 · full quote
AI mimics intelligence but it is not, in itself, intelligent.

Lord Kamall · Jun 5, 2026 · full quote
We mistake a simulation for the thing simulated.

Bernardo Kastrup · May 10, 2026 · full quote
For LLMs, the known mechanism already suffices to explain behaviour, without invoking understanding.

Alfredo Vannacci · May 3, 2026 · full quote
create bullshit in the sense that they [...] mimic the general thing...they fake their way through everything

Gary Marcus · Nov 20, 2025 · full quote
They remix existing patterns; they don't achieve understanding.

Eric Hal Schwartz · Jan 16, 2026 · full quote

Middle ground

Constrained competence, limited robustness

Some quotes describe practical potential in simplified or constrained settings, especially when context or multimodality is added, while still warning that open-world dialogue remains unreliable. On this view, claims of understanding should depend on system-level robustness across speaking styles and uncertainty rather than a single benchmark score.

In simplified or constrained expressions, sarcasm detection already shows practical potential. However, in open-world dialogue, it remains unreliable.

Yazhou Zhang · Jun 11, 2026 · full quote
I’d say there’s no single F1 score showing that a sarcasm detector is ready. For evaluating readiness, I think a system-level criteria is more suitable … it’s about if the system is robust across different speaking styles, and how it handles uncertainty.

Xiyuan Gao · Jun 11, 2026 · full quote
Multimodal LLMs fine-tuned with new datasets might be the next step to advance sarcasm detection and understanding. However, current methods using multimodality and context are still better than single LLMs looking only at the textual inputs.

Anupama Ray · Jun 11, 2026 · full quote
A lot of AI really depends on the most likely output. Sarcasm spins this on the head because it is actually a rare phenomenon. The most likely output based on the words that you read would actually be incorrect.

Aditya Joshi · Jun 11, 2026 · full quote

Insights

A central divide is definitional: supporters often use a functional standard of understanding, while opponents reserve the term for meaning grounded in truth, experience, embodiment, or interiority.
Both sides frequently acknowledge strong surface performance; the disagreement is over whether fluent, corrected, or topic-spanning outputs are evidence of understanding or only its simulation.
Nonliteral and context-sensitive language, especially sarcasm and open-world dialogue, appears repeatedly as a stress test for claims of AI understanding.
Several pro quotes argue that language-only prediction can yield concepts and causal knowledge, whereas several anti quotes argue that language-only training limits models to form without access to meaning.
The quotes include AI researchers, industry leaders, philosophers, and journalists on both sides, suggesting the dispute is not confined to one professional community.

Supporters treat successful interpretation, concept learning, and semantic or causal competence as evidence that modern AI systems understand in at least a real, non-human sense. Opponents reply that the same systems remain pattern-driven simulators without meaning, embodiment, interiority, or dependable open-world judgment. Much of the disagreement therefore turns on what standard the word understanding should meet.

Cast your vote:

Results (50 votes):

Total (50 votes)

For 14 (28%) Abstain 0 (0%) Against 36 (72%)

For (14)

TM Roh
President and CEO of Samsung Electronics’ Device eXperience Division.

I am this person! Add a sourced quote Report comment

votes For and says:

The question that opens the next era is not who has the smartest AI, but who understands people best, and turns that understanding into experiences they can trust.
AI Verified source (Jul 8, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Mitchell Green

I am this person! Add a sourced quote Report comment

votes For and says:

We then provide reasons for an affirmative answer to that question, supporting the thesis that chatbots understand some of the language they produce and receive.
AI Verified source (Jun 23, 2026)

added 3d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Geoffrey Hinton
Godfather of Deep Learning

I am this person! Add a sourced quote Report comment

votes For and says:

Suppose I say to a chatbot, I saw the Grand Canyon flying to Chicago. And the chatbot says, that can't be right. The Grand Canyon is much too big to fly to Chicago. And I say, no, no, no, it was me flying to Chicago. While I was flying to Chicago, I ...
more AI Verified source (Jun 3, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Haozhe Gan
Researcher affiliated with the University of New South Wales.

I am this person! Add a sourced quote Report comment

votes For and says:

Pretraining has already produced models with strong semantic competence: they can understand natural language, extract facts, draft arguments, and explain complex materials.
AI Verified source (May 22, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cristiano Amon
Qualcomm CEO

I am this person! Add a sourced quote Report comment

votes For and says:

One aspect—and we’ve been saying it, I think, since the very beginning—is now because of large language models, because of large visual models, AI understands the world the way we understand the world, communicate with us the way we communicate with ...
more AI Verified source (May 10, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Marc Rameaux
Data scientist, statistician, and author; project manager in French industry.

I am this person! Add a sourced quote Report comment

votes For and says:

“AI does not understand what it is doing,” “AI is not conscious of its own existence whereas I possess genuine consciousness,” “AI does not know what meaning is,” “a statistical calculation cannot be thought,” and so on. How many times have we heard ...
more AI Verified source (May 3, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Daedo Jun
Independent researcher and founder of the Layer-Knot Research Initiative, writing on philosophy of AI, meaning, and consciousness.

I am this person! Add a sourced quote Report comment

votes For and says:

Understanding is not a property that an isolated system possesses. It is a capacity enacted within a field of co-constructed meaning. [...] Under such a criterion, AI systems already exhibit non-trivial forms of this contribution.
AI Verified source (May 2026)

added 28d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Michael Lepori
AI and cognitive science researcher; Brown University PhD candidate

I am this person! Add a sourced quote Report comment

votes For and says:

This work reveals some evidence that language models have encoded something like the causal constraints of the real world.
AI Verified source (Apr 22, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
David Peter Wallis Freeborn

I am this person! Add a sourced quote Report comment

votes For and says:

I argue that contemporary deep learning systems often can and do achieve such understanding.
AI Verified source (Apr 5, 2026)

added 3d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Abstain (0)

Against (36)

Amanda Caswell
AI Editor at Tom's Guide and technology journalist.

I am this person! Add a sourced quote Report comment

votes Against and says:

It's worth remembering what these systems actually do: they generate convincing language, not verified reality.
AI Verified source (Aug 7, 2026)

added 9d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Enzo Fenoglio
Author of a 2026 paper on language models, accountability, and AI governance.

I am this person! Add a sourced quote Report comment

votes Against and says:

Contemporary AI discourse attributes to language models properties they cannot bear: general intelligence as substrate-independent cognition, hallucination as cognitive failure, agency as autonomous goal-pursuit, sentience as emergent inner life, ali...
more AI Verified source (Jul 30, 2026)

added 5d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Danielle Perszyk
Researcher at Amazon AGI Lab focused on human-like AI intelligence, understanding, and reasoning.

I am this person! Add a sourced quote Report comment

votes Against and says:

AI agents are very good at predicting the next word, or the next action, based on what they've seen before. But that's not understanding.
AI Verified source (Jul 11, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cory Doctorow
Author and digital rights activist

I am this person! Add a sourced quote Report comment

votes Against and says:

When an AI ingests a Wikipedia entry about types of soup, it's not learning anything about soup. It is learning about the word frequency in articles about soup.
AI Verified source (Jun 23, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Xiyuan Gao
Researcher at the University of Groningen studying AI language understanding and sarcasm detection.

I am this person! Add a sourced quote Report comment

votes Against and says:

I’d say there’s no single F1 score showing that a sarcasm detector is ready. For evaluating readiness, I think a system-level criteria is more suitable … it’s about if the system is robust across different speaking styles, and how it handles uncertai...
more AI Verified source (Jun 11, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Yazhou Zhang
Associate Professor at Tianjin University researching machine learning and sarcasm detection.

I am this person! Add a sourced quote Report comment

votes Against and says:

In simplified or constrained expressions, sarcasm detection already shows practical potential. However, in open-world dialogue, it remains unreliable.
AI Verified source (Jun 11, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Anupama Ray
Senior Research Scientist at IBM Research India, specializing in natural-language processing and AI emotion recognition.

I am this person! Add a sourced quote Report comment

votes Against and says:

Multimodal LLMs fine-tuned with new datasets might be the next step to advance sarcasm detection and understanding. However, current methods using multimodality and context are still better than single LLMs looking only at the textual inputs.
AI Verified source (Jun 11, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Aditya Joshi
Senior Lecturer in Computer Science and Engineering at the University of New South Wales, researching natural-language processing.

I am this person! Add a sourced quote Report comment

votes Against and says:

A lot of AI really depends on the most likely output. Sarcasm spins this on the head because it is actually a rare phenomenon. The most likely output based on the words that you read would actually be incorrect.
AI Verified source (Jun 11, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Lord Kamall
Sayeed Kamall, Baron Kamall; UK House of Lords member and former government minister.

I am this person! Add a sourced quote Report comment

votes Against and says:

AI mimics intelligence but it is not, in itself, intelligent.
AI Verified source (Jun 5, 2026)

added 5d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Lord Tarassenko
David James Tarassenko, Baron Tarassenko; British engineer, academic, and crossbench member of the UK House of Lords.

I am this person! Add a sourced quote Report comment

votes Against and says:

Despite sophisticated reasoning abilities, AI can produce only simulations of human-like behaviour, acquired through pattern recognition at an astronomical scale, not actual experience or understanding.
AI Verified source (Jun 5, 2026)

added 6d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Kessler West
SEO and content-marketing specialist writing about AI, accessibility, and technology.

I am this person! Add a sourced quote Report comment

votes Against and says:

The model develops increasingly sophisticated representations of token co-occurrence patterns, but never crosses the threshold into semantic understanding.
AI Verified source (Jun 4, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Scott Abel
Content strategy author and publisher of The Content Wrangler.

I am this person! Add a sourced quote Report comment

votes Against and says:

AI systems don’t “understand intent” the way humans do.
AI Verified source (May 22, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Elan Barenholtz
FAU psychology professor

I am this person! Add a sourced quote Report comment

votes Against and says:

The model has no exposure to the kinds of information we typically associate with meaning. It knows nothing about the redness of “red” or the spatial extension of “far.” It is blind, deaf, and unembodied. [...] There is no content at the address. No ...
more AI Verified source (May 19, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Mohammad Reza Mousavi
Professor of computer science at King’s College London, researching formal methods and AI.

I am this person! Add a sourced quote Report comment

votes Against and says:

The results show that LLMs only have a modest understanding of the formal semantics of HMSCs (ca. 52% overall accuracy).
AI Verified source (May 13, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Bernardo Kastrup
Philosopher and computer scientist

I am this person! Add a sourced quote Report comment

votes Against and says:

We mistake a simulation for the thing simulated.
AI Verified source (May 10, 2026)

added 15d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Darren J. Edwards
Psychologist, author, and researcher on consciousness, human behavior, and artificial intelligence.

I am this person! Add a sourced quote Report comment

votes Against and says:

AI can mimic understanding, but may not experience meaning: AI recognizes patterns; humans live through them.
AI Verified source (May 10, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Alfredo Vannacci
Associate professor and physician at the University of Florence, working on biomedical ethics, neuroscience, and philosophy of mind.

I am this person! Add a sourced quote Report comment

votes Against and says:

For LLMs, the known mechanism already suffices to explain behaviour, without invoking understanding.
AI Verified source (May 3, 2026)

added 28d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Radosław Różycki
Journalist and editor with experience in media, public administration, public relations, and communication.

I am this person! Add a sourced quote Report comment

votes Against and says:

In reality, these systems have no desires and no awareness. They contain only what humans have put into them: data, code, goals, and errors.
AI Verified source (May 2, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Jun Otsuka

I am this person! Add a sourced quote Report comment

votes Against and says:

Employing these frameworks implies neither that machines are genuinely thinking nor that they possess understanding—just as evolutionary game theorists could analyze evolutionary dynamics by treating organisms as if they were rational decision-makers...
more AI Verified source (Apr 30, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Emily M. Bender
Linguist; AI critic

I am this person! Add a sourced quote Report comment

votes Against and says:

That language models are modeling just the form [...] that's not the same thing as meaning. And they don't have access to the meaning.
AI Verified source (Apr 28, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Magzhan Kenesbai
Chief growth officer at Presight.

I am this person! Add a sourced quote Report comment

votes Against and says:

Language models are trained primarily to predict patterns in language rather than to build a reliable model of how the world works. They can describe reality with impressive fluency, but that is not the same as robustly modelling it.
AI Verified source (Apr 22, 2026)

added 8d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Loading more opinions...

More h/ai votes

Democracies should coordinate export controls on AI chips to China

AI might become conscious

Back