The question of whether AI can “understand” humans depends greatly on what we mean by “understanding.”

If we define it narrowly—predicting human behavior, inferring intentions, or interpreting emotions—then recent advances in Large Language Models (LLMs) suggest that AI can simulate some aspects of human understanding through what is known in cognitive science as Theory of Mind (ToM).

Theory of Mind refers to the ability to attribute mental states—beliefs, desires, intentions—to others and to predict their behavior based on those states. Traditionally, this is seen as a uniquely human (and partially animal) capability, acquired through social interaction and embodied experience.

Recent studies have tested LLMs with classical ToM tasks, such as false-belief reasoning. For example, Kosinski (2023) reported that GPT-3.5 and GPT-4 solved ToM tasks at a level comparable to or exceeding that of adult humans (“Theory of Mind May Have Spontaneously Emerged in Large Language Models,” arXiv:2302.02083). Other works, such as Ullman (2023), have challenged these findings, suggesting that LLM performance may rely on pattern recognition in training data rather than genuine mental-state representation (“Large Language Models Fail on Trivial Alterations to Theory-of-Mind Tasks,” arXiv:2302.08399).

This raises the crucial distinction between performance and genuine understanding. Passing a ToM benchmark does not necessarily mean an AI experiences or introspects on human feelings and intentions—it may simply reproduce statistical regularities from language data. Moreover, AI lacks embodied experience and first-person perspective, which in humans are thought to ground empathy, self-awareness, and moral reasoning.

That said, with the integration of multimodal architectures (language, vision, audio, sensorimotor data) and training in interactive environments, AI systems could develop increasingly functional models of human mental states. Whether this will ever amount to the kind of rich, experiential “understanding” humans attribute to one another remains an open philosophical and scientific question.

Dear Takao Inoue ,

No, machines lack subjective experiences like emotions. They can simulate responses but don't feel. What is the main limitation of AI understanding? AI lacks personal experiences, making its understanding surface-level and abstract.

https://www.reddit.com/r/ArtificialInteligence/comments/12pec52/can_ai_actually_understand_us_and_what_does_it/

The transformative achievements of deep learning have led several scholars to raise the question of whether artificial intelligence (AI) can reach and then surpass the level of human thought. Here, after addressing methodological problems regarding the possible answer to this question, it is argued that the definition of intelligence proposed by proponents of the AI as “the ability to accomplish complex goals,” is appropriate for machines but does not capture the essence of human thought. After discussing the differences regarding understanding between machines and the brain, as well as the importance of subjective experiences, it is emphasized that most proponents of the eventual superiority of AI ignore the importance of the body proper on the brain, the laterization of the brain, and the vital role of the glia cells. By appealing to the incompleteness theorem of Gödel’s and to the analogous result of Turing regarding computations, it is noted that consciousness is much richer than both mathematics and computations.

Article Can artificial intelligence reach human thought?

Regards,

Shafagat

AI Understands Differently Than Humans

AI can process, predict, and respond in ways that approximate human understanding, but it does not experience meaning, emotion, or intent. For now, AI is a powerful tool for simulating understanding!! not replicating it. The deeper question isn't just whether AI understands us, but whether we’re comfortable calling statistical pattern recognition "understanding" at all.

1. What Does It Mean for a Human to Be Understood?

Human understanding is a multi-layered phenomenon involving:

• Cognitive Comprehension : Grasping the literal meaning of words, concepts, or instructions.
• Emotional Intelligence : Recognizing tone, intent, sarcasm, humor, or unspoken feelings.
• Contextual Awareness : Interpreting meaning based on cultural, social, and situational factors.
• Empathy & Theory of Mind : Inferring beliefs, desires, and perspectives beyond explicit communication.

A human is "understood" when another being (human or artificial) can not only process their words but also interpret their deeper meaning, respond appropriately, and predict their needs or reactions.

2. Can AI Achieve This Level of Understanding?

Current AI, particularly Large Language Models (LLMs), exhibits partial understanding in specific ways:

1. Semantic Processing : AI can parse language, summarize text, and generate coherent responses. 2. Pattern Recognition : It detects statistical relationships in data, allowing it to mimic human-like dialogue. 3. Contextual Adaptation : Advanced models retain conversation history and adjust responses accordingly.

Meanwhile; AI fundamentally lacks:

• Subjective Experience : No consciousness, emotions, or self-awareness.
• True Intentionality : It predicts text but does not "mean" what it says.
• Embodied Cognition : Humans understand through lived experience; AI has no body, no senses, no personal history.

3. The Future: Will AI Ever Truly Understand?

• Narrow AI (today's systems) can simulate understanding within limited domains (e.g., chatbots, recommendation systems).
• Artificial General Intelligence (AGI) : if achieved; might replicate deeper comprehension, but this remains speculative.
• Philosophical Limits : Some argue that understanding requires consciousness, which AI may never possess.

AI can detect human sentiments.

From my experience a user, AI looks as if understanding humans when it aligns itself with the human feelings /emotions, and identifies with them empathically, on top of being correct with its answer.