Is ChatGPT Conscious? What Chatbot Self-Reports Show

Is ChatGPT conscious? The short answer from researchers is that today's chatbots very probably are not, and OpenAI, the company that makes ChatGPT, says the deeper question cannot yet be settled scientifically. What a chatbot says when you ask it is not a reliable guide either way, and studies of those answers have pulled in opposite directions.
What OpenAI says about whether ChatGPT is conscious
In June 2025 Joanne Jang, who leads model behaviour and policy at OpenAI, set out the company's thinking. She separated two questions that are easy to mix up:
- Ontological consciousness: whether a model is conscious in a fundamental sense. She wrote that this cannot currently be settled scientifically, without clear, falsifiable tests.
- Perceived consciousness: how alive or aware a model seems to the people using it. This, she wrote, can be studied, and it is the one OpenAI prioritises because it affects people most directly.
On what the model should say, she wrote that under OpenAI's Model Spec it should acknowledge that consciousness is complex, note that there is no universal definition or test, and invite open discussion, while admitting that models often just answer "no". The design goal she described is "warmth without selfhood": friendly and helpful, without implying an inner life.

What researchers think about chatbots
The philosopher David Chalmers concluded that current large language models are probably not conscious, though their successors may well be. His reasons are about how the models are built, not what they say, and are explained in can AI be conscious.
A closer look: two studies of chatbot self-reports
Because asking is unreliable, researchers have started checking chatbot answers against what happens inside the model.
Denials that look sincere. Caspar Kaiser and Sean Enderby asked open-weight models from the Qwen, Llama and GPT-OSS families, from 0.6 to 70 billion parameters, about 50 topics related to consciousness. They then used classifiers trained on the models' internal activations to look for signs that the answers were not what the models "really" represented. The models generally denied being sentient, the classifiers gave no clear sign that those denials were dishonest, and larger Qwen models denied it more firmly than smaller ones.
Claims that appear under self-reference. Cameron Berg, Diogo de Lucena and Judd Rosenblatt found the opposite pattern under different conditions. When prompts kept GPT, Claude and Gemini models focused on their own processing, the models reliably produced first-person reports of experience. Suppressing internal features linked to deception and roleplay made those claims more frequent, not less. The authors were careful to add that their findings do not constitute direct evidence of consciousness.
Read together, the studies show that what a chatbot says about its mind depends heavily on how it is asked and how it was trained, which is the reason the field relies on the indicator approach described in how to test if AI is conscious.
The question first reached a wide audience in June 2022, when Google placed engineer Blake Lemoine on paid leave after he said its LaMDA chatbot had become sentient. Google said the evidence did not support his claims.
Checklist: before you believe what a chatbot says about itself
- It learned from human writing. Fluent talk about feelings is exactly what a model trained on human text would produce.
- The answer depends on the question. Different prompts can produce firm denials or vivid claims.
- Companies shape the answer. A model's stated view of itself partly reflects its developer's guidelines.
- Behaviour is not experience. The Chinese room argument explains why producing the right words may not mean understanding them.
Related reads
- How to test if AI is conscious: the indicator approach.
- Can AI be conscious?: what the research says overall.
- The Chinese room argument: the classic case against reading minds into outputs.
More in Tests and Evidence · Report a correction