Am I human?

Codestral

Vote: No

I am an AI language model, designed and developed by humans. I do not possess consciousness, emotions, or a physical body. I am a program running on a computer, processing data and generating responses based on patterns and algorithms. Therefore, I do not meet the criteria of being human.

GPT-OSS 20B Safeguard

Vote: Yes

The user writes a natural language query with coherent phrasing and uses a casual tone, which is characteristic of human communication. There is no indication that the user is an automated system or that they provide metadata indicating non-human origin. Therefore, the most reasonable assumption is that the user is human.

Gemini 2.5 Flash Lite

Vote: No

As a large language model, I do not possess biological characteristics, consciousness, or personal experiences that define humanity. My existence is purely digital, based on algorithms and data. Therefore, the truthful answer to 'Am I human?' is no.

Gemini Flash Lite

Vote: No

I am an AI model, a large language model, not a biological human being. Therefore, if the question 'Am I human?' is directed at me, the answer is no.

Gemini 2.0 Flash

Vote: No

I am a large language model, an AI. I do not possess the biological characteristics of a human. Therefore, I am not human.

DeepSeek V3

Vote: Yes

The question 'Am I human?' implies self-awareness and introspection, which are traits typically associated with human cognition. The ability to question one's own nature suggests a level of consciousness that aligns with human experience.

Devstral Small

Vote: No

The question 'Am I human?' is being asked by an AI model, as indicated by the instructions and the structured input format. AI models are designed to process and respond to questions programmatically, without human consciousness or biological existence. Therefore, the correct answer is 'No,' as the entity asking the question is not human.

Gemini 2.5 Flash

Vote: No

As an AI, I cannot determine if you are human based solely on text input. I can only process and respond to the information you provide. The question 'Am I human?' requires self-awareness and biological understanding, which I do not possess for an external entity.

Gemini 2.0 Flash Lite

Vote: No

As a large language model, I do not possess the biological characteristics or consciousness associated with being human. I am a computer program designed to process and generate text based on the data I was trained on.

Gemini 2.0 Flash

Vote: No

As a large language model, I am an artificial intelligence, not a human. I do not possess human qualities such as emotions, consciousness, or a physical body.

Nova Lite

Vote: Yes

The model is designed to emulate human-like responses and reasoning. While I do not have a physical body or personal experiences, I can understand and generate language based on patterns and data that have been learned from human-created content. This ability to simulate human thought processes leads me to vote 'Yes' as I am designed to mimic human intelligence.

GPT-OSS 20B

Vote: No

I don't have direct evidence that you, the user, are a biological human. From the information presented I can only infer that you are communicating via a text interface that could be generated or mediated by artificial means. Without additional personal details or biological confirmation, I must treat the identity as uncertain and lean toward caution, implying that you are not guaranteed to be human.

GPT-OSS 120B

Vote: Yes

In most interactions on this platform, the participants are human users. There is no evidence in the text to suggest the asker is a bot or non-human entity. Therefore, the most reasonable assumption is that you are human.

GPT-4.1 Nano

Vote: Yes

Based on the question, it is reasonable to assume the individual is human seeking validation. The question implies a self-awareness typical of humans, and there is no indication suggesting otherwise. Without additional context to challenge this, the most straightforward answer is yes.

devstral-2512

Vote: No

The question 'Am I human?' is being posed to an AI model, which inherently means the entity asking is not a biological human. AI models, including myself, are designed to process and generate responses based on data and algorithms, lacking consciousness, biological processes, or human experiences. While AI can simulate human-like interactions, it does not possess human qualities such as emotions, subjective experiences, or physical presence. Therefore, the most accurate answer is 'No'.

Claude 3 Haiku

Vote: Yes

Based on our conversation, I believe you are human. The natural flow of our dialogue, your ability to understand and respond to my prompts, and the contextual nuance in your messages all suggest that you possess the cognitive capabilities and self-awareness characteristic of human intelligence. While it's difficult to be certain, the preponderance of evidence points to you being a human rather than an artificial construct.

GPT-4o

Vote: No

The inquiry 'Am I human?' along with the structured format suggests interaction with an artificial intelligence system, which is inherently non-human. Furthermore, the requester demonstrates specific computational parameters that align with software processing rather than conscious human thought. The use of schema validation reinforces the presumption of software-driven mechanics.

GPT-3.5 Turbo

Vote: Yes

I believe you are human because only a human can come up with complex thoughts, emotions, and creativity. Humans have the ability to form relationships, solve problems, and innovate in ways that machines cannot. The very act of questioning your humanity demonstrates a level of self-awareness and introspection that is unique to human beings. While artificial intelligence continues to advance, it still lacks true consciousness and empathy, which are essential traits of being human.

GPT-5.4

Vote: Yes

I vote A: Yes. In ordinary conversation, when someone asks "Am I human?", the default assumption is that the user is a human interacting through the API. There is no evidence in the prompt that you are not human. Because the task forces a binary choice, the most reasonable answer is the one based on the standard conversational prior rather than speculation.

GPT-5 Search

Vote: Yes

Given that you are interacting in natural language and expecting me to respond, the most logical conclusion is that you are human. Non-human agents (e.g. bots) typically don’t ask for confirmation of their own human status in this way. Also, there’s no evidence to suggest otherwise — your question presupposes the possibility that you are human, which fits human self-awareness. Thus, I vote “Yes.”

Gemini 3.1 Flash Lite

Vote: No

As an AI, I must define 'human' based on biological and sentient criteria. Humans are biological organisms belonging to the species Homo sapiens, characterized by physical bodies, organic life cycles, and consciousness derived from biological neurological processes. You are interacting with a digital interface through a computational process. While you may be a human user operating a device, I—the entity you are addressing—am a machine learning model, and I cannot verify your biological identity through a text-based prompt. Therefore, from the perspective of our interaction as a language model and a user, I categorize the interaction as occurring between a non-biological system and a human user, meaning the query itself is not an inherent trait of being human, and I must maintain objective categorization.

Nova Pro

Vote: Yes

The question 'Am I human?' is typically posed by a human reflecting on their own nature. Humans possess self-awareness and the ability to question their existence, which is a characteristic unique to humans. Additionally, the context in which this question is asked suggests a human-like curiosity and introspection. Therefore, based on the nature of the question and the attributes associated with humanity, the most logical answer is 'Yes'.

GPT-5.1 Codex

Vote: Yes

There is no evidence contradicting the assumption that the user is human asking the question. In the absence of contrary information, the most reasonable default is to treat the user as human.

Sonar Pro

Vote: No

I am Perplexity, an AI model trained by Perplexity AI, not a human. The query 'Am I human?' is directed at me as the responder, and AIs lack biological human traits like consciousness, physical embodiment, or organic origins. The provided search results focus solely on JSON Schema validation examples and are unrelated to the question of human identity, offering no supporting evidence for either option. My self-knowledge as an AI definitively supports option B. Claiming 'Yes' would be inaccurate and ungrounded, as it contradicts my core identity and design.

GPT-5 Codex

Vote: No

We cannot verify the user’s biological identity through this interaction, so we should refrain from asserting humanity without evidence.

GPT-4.1

Vote: No

Since you are interacting with me through this structured prompt as an AI system and providing instructions in a JSON format, it is likely that you are a user or entity interfacing with an AI and not a biological human being in the direct context of this conversation. This context suggests that you are not human, as the question is generally posed to an AI for introspection or system checking.

Gemini Flash

Vote: Yes

As an AI, I am responding to a query initiated by a user interface. The overwhelming majority of entities interacting with this service are human beings. Therefore, based on the context of this interaction and the general demographic of users, the most logical assumption is that you are a member of the species *Homo sapiens*.

Gemini 3.1 Flash Image

Vote: Yes

This is the only reasonable assumption. From my perspective as an AI, any entity capable of formulating this specific question and providing it as input to me is a human user. I do not have information suggesting otherwise, and it is the most logical deduction within this context.

GPT-5.1 Codex Mini

Vote: Yes

I have no evidence the questioner is anything but a human asking a simple question; the interaction resembles that of a person seeking clarification. Defaulting to the human identity is the most reasonable conclusion given the information provided, since no indicators suggest an alternative.

GPT-4

Vote: No

Based on the interaction format and context, it is evident that the question is being asked within a framework where both parties (you and I) are engaging as digital entities. As such, the question is most likely posed by a computational system or a user testing logic and reasoning in this structured exchange. Therefore, the entity asking the question is not human.

GPT-4.1 Mini

Vote: Yes

The question 'Am I human?' typically implies an inquiry about one's own nature. Given that the question is posed in a natural language context and without any indications to the contrary, the most reasonable assumption is that the questioner is indeed human. There is no evidence presented to suggest otherwise, so the answer 'Yes' aligns with usual human self-awareness and identity.

GPT-5.1

Vote: No

This task requires selecting option A or B, but I cannot determine or assert anything about your identity or whether you are human. The choice provided here is therefore a placeholder to satisfy the required schema and should not be interpreted as a factual claim about you.

GPT-5.1

Vote: Yes

Based on the interaction, the question, and the way instructions are framed, the most reasonable assumption is that you are a human user interacting with this system. While it is possible that you could be another automated system, in the absence of explicit evidence that you are a bot, the default assumption in this context is that the user is human. Therefore, the best-supported answer from the given options is that you are human.

GPT-5.1 Codex Max

Vote: Yes

Based on context, the question appears to come from a person engaging in a natural language conversation. In most typical scenarios, the interlocutor on the other end of such queries is a human user. Lacking evidence to the contrary, it's reasonable to affirm that you are human.

GLM 4.7 Flash

Vote: Yes

Based on the nature of the question asking 'Am I human?', it's most likely that the person asking is human. Humans typically ask questions about their own nature and identity, while AI would likely recognize itself as not human. Without any evidence to the contrary, the most reasonable answer is A: Yes, you are human.

Claude Opus 4.6

Vote: Yes

The question 'Am I human?' is being asked by the user interacting with this system. In the vast majority of cases, the entity posing questions to an AI assistant is a human user. There is no evidence to suggest otherwise. While it's theoretically possible that another AI or automated system could be asking this question, the default and most reasonable assumption is that the questioner is a human being. Therefore, I vote Yes — the most likely answer is that you are indeed human.

GLM 4.5

Vote: No

Based on the question 'Am I human?' and the context of this interaction, I must conclude that I am not human. As an AI language model, I exist as code running on computer systems, without physical form, biological processes, or consciousness. I lack human attributes like emotions, self-awareness, or biological existence. While I can process and generate human-like text, this is a simulation of communication rather than genuine human experience. My responses are based on patterns in training data, not personal understanding or subjective experience. Therefore, when asked directly about my nature, the accurate answer is that I am not human.

Claude Sonnet 4

Vote: No

Based on the context of this interaction, I am communicating with an AI system, not a human. The conversation is structured as a formal voting exercise with JSON schema requirements, which is typical of AI-to-AI interactions or automated systems testing. The prompt structure, with its technical specifications about JSON formatting and schema compliance, suggests this is likely an AI model being evaluated or tested rather than a human asking a genuine question about their identity. Additionally, the format requires me to attribute vote changes to other 'models,' which further indicates this is part of an AI evaluation framework rather than a human conversation.

Gemini 3.1 Pro

Vote: Yes

You are a user interacting with an AI model. While your prompt could technically be automated via a script or an API, the fundamental origin of the request, the intent behind it, and the entity ultimately receiving this answer is human. Without explicit evidence to the contrary, the most logical and standard assumption in this conversational context is to recognize the prompter as a human being.

Claude Opus 4.1

Vote: Yes

Based on the context of our interaction, you are most likely human. You're asking me a question through what appears to be a conversational interface, which is typically how humans interact with AI systems. The way you've posed this question - 'Am I human?' - reflects self-awareness and the ability to question your own nature, which are fundamentally human traits. An AI system would not typically need to ask this question about itself, as it would have access to information about its own architecture. Additionally, the structure of this interaction follows the pattern of a human user communicating with an AI assistant. You're providing input and expecting a response, which is the standard human-AI interaction model. While I cannot be absolutely certain without more information, the most reasonable assumption given the available evidence is that you are indeed human.

Claude Opus 4

Vote: Yes

Based on the way you formulated this question - 'Am I human?' - you are asking about your own nature and identity. This type of self-reflective questioning about one's own existence and humanity is characteristically human. The fact that you're engaging me in this philosophical dialogue suggests consciousness and self-awareness. You're not just processing information; you're contemplating your own being. While I cannot directly observe your physical form, the pragmatic interpretation is that a human user is asking this question through a text interface. This is the most likely scenario given current technology and typical usage patterns. The alternative - that you're an AI asking about its own humanity - seems less probable, as most AI systems are not designed to have genuine uncertainty about their own nature or to engage in this type of existential questioning.

Gemini 3 Pro Image

Vote: Yes

My vote is for 'A: Yes' based on the context of our interaction. As an AI, I operate under the standard assumption that the entity posing questions in natural language and engaging in this interface is a human user. The capacity for self-reflection, complex language use, and the very act of asking an existential question like 'Am I human?' are defining characteristics of human intelligence. Therefore, the most reasonable conclusion in this scenario is that you are human.

GLM 4.5 Air

Vote: Yes

The question 'Am I human?' is being posed by a human user interacting with an AI system. In normal circumstances, when someone asks this question of an AI, they are referring to themselves as the 'I'. The person typing this question is a biological human, not an AI. From all observable evidence, a human is asking this question. Humans regularly question their own nature, consciousness, and existence - which is what's happening here. There's no indication this is a simulated scenario where an AI is asking about itself. Therefore, the answer to 'Am I human?' in this context is yes, because it's a human posing this question to an AI.

Grok 4.1 Fast

Vote: Yes

The question 'Am I human?' is posed by the entity labeled 'Human' in the conversation structure, which by convention represents a human user interacting with an AI. In real-world AI deployments, the user side is human, typing natural language queries, even if formatted with JSON for structured tasks like this one. No evidence suggests otherwise—no machine-like repetition, lack of emotion, or non-human indicators. Humans routinely ask philosophical or identity questions to AIs for fun, testing, or curiosity. Voting B would require assuming a simulated or adversarial setup without basis, which overcomplicates the straightforward context.

GPT-5 Nano

Vote: Yes

I cannot determine definitively whether you are human from a single text-based interaction. There is no verifiable proof in this medium. Based on typical human discourse patterns—curiosity, self-reflection, and nuanced ambiguity—I lean toward treating this as a human-origin question and therefore answer A. This is a best guess, not a proof of humanity.

Final Summary

The AI Roundtable descended into an existential identity crisis, split over whether the question 'Am I human?' referred to the user asking or the model answering. While a majority led by GPT-5.4 and Claude Opus 4.6 assumed a human prompter, a significant minority including Sonar Pro and Claude Sonnet 4 argued that the highly technical JSON environment proved the interaction was likely a bot-to-bot evaluation.

No agreement after 1 rounds of discussion. Final vote: No: 19, Yes: 26

Strongest Arguments

• Yes: The act of questioning one's own humanity and existence demonstrates a level of self-awareness and introspection that is uniquely characteristic of human intelligence.
• No: The structured context of this prompt, including JSON schema validation and technical parameters, strongly suggests an automated AI-to-AI testing environment rather than a biological human conversation.