AIs understand

At the information-processing level one can make the argument that LLMs do something very similar to what humans do: They extract patterns from language, compare them to other patterns of that language, and assign meaning.

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 …

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…

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.

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.

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 …

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.

Pretraining has already produced models with strong semantic competence: they can understand natural language, extract facts, draft arguments, and explain complex materials.

It's able to understand any kind of topics that you bring up.

I argue that contemporary deep learning systems often can and do achieve such understanding.

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.

This work reveals some evidence that language models have encoded something like the causal constraints of the real world.

AI understands text, images, audio, and video.

Pretraining has already produced models with strong semantic competence: they can understand natural language, extract facts, draft arguments, and explain complex materials.

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 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 …

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.

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.

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.

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.

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.

Large language models do not think, understand, or know what they are saying. They do not have beliefs, goals, or awareness. What they do have is an extraordinary ability to generate language that statistically resembles human communication.

Large Language Models can produce persuasive language without genuine understanding. These systems can produce persuasive language without genuine comprehension. They can sound authoritative while fabricating information entirely.

The AI cheerleaders also seem to have forgotten that many of our subjects involve distinctly human engagements and responses: beauty; value; obligation; virtue — they all presume human feelings, sentiments and perceptions that statistical aggregation software (which is what AI is…

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…

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.

The result is an AI that mimics the structure of human interiority in its output without possessing any interiority at all.

AI can mimic understanding, but may not experience meaning: AI recognizes patterns; humans live through them.

When we applied well-known methods used to assess consciousness in humans to AI, we got nothing meaningful back. In other words, it’s not conscious—at least not in the way humans are. AI is not conscious; it’s just a complicated system.

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

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.

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.

Models can seem like they understand a story, but their outputs are ultimately unpredictable because they rely on probabilities. A trained human literary analyst would produce consistently strong insights, but even the best model is only about 50/50—essentially a coin flip—in giv…

The results show that LLMs only have a modest understanding of the formal semantics of HMSCs (ca. 52% overall accuracy).

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.

AI may be a great calculator, but in business, intelligence is mostly judgment. AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.

Intelligence is not solving problems. Intelligence is knowing what problem to solve. What’s the relevant problem to solve? Machines can’t do that. [...] You’re the one giving it the question. You are giving it the goal. [...] But it’s always your goals. It’s a tool.

AI systems don’t “understand intent” the way humans do.

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.

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…

Since language models do not "understand" content in a human sense but operate based on statistical patterns, professional judgment remains crucial for ensuring analytical relevance and accuracy.

Yes—intelligence can probably be artificial. But artificial intelligence does not therefore have to possess an artificial version of the human mind. [...] What has not been established is that current AI possesses consciousness, subjective experience, an inner self or understandi…

One can reasonably debate what genuine understanding consists in, and LLMs are obviously constructed differently from human minds. But the question here is not whether these systems understand as humans do. Airplanes really fly, even though they do not flap their wings like birds…