It\u2019s a position of highly suggestive uncertainty. Anthropic is effectively telling people that it believes that chatbots might already be thinking, feeling entities \u2014 far more publicly than OpenAI, xAI, Google, or virtually any other major consumer AI company. It\u2019s making a claim many<\/a> experts<\/a> conclude<\/a> is an extreme long shot, while reinforcing ideas that have caused real harm<\/a>, including some deaths by suicide among people who believe that the chatbot they\u2019re speaking with exhibits some form of consciousness or deep empathy.<\/p>\n Over the course of interviews for podcasts<\/a>, profiles<\/a>, and feature articles<\/a>, Amodei and other company leaders have repeatedly refused to rule out the possibility that Claude might be conscious and instead raised questions about how something can be conscious in a different way than humans. Anthropic\u2019s chief philosopher Amanda Askell told The New Yorker, \u201cIf it\u2019s genuinely hard for humans to wrap their heads around the idea that this is neither a robot nor a human but actually an entirely new entity, imagine how hard it is for the models themselves to understand it!\u201d<\/p>\n These interviews don\u2019t precisely define \u201cconscious,\u201d a term whose meaning experts disagree on anyway. A starting point, from the Merriam-Webster dictionary, is \u201cthe quality or state of being aware especially of something within oneself\u201d or \u201cthe state of being characterized by sensation, emotion, volition, and thought.\u201d That seems not far off from Anthropic\u2019s use of the term.<\/p>\n Anthropic is \u201cnot even sure that we know what it would mean for a model to be conscious \u2026 But we\u2019re open to the idea that it could be.\u201d<\/p>\n Many scientists say it\u2019s not possible for AI systems like large language models to become conscious in any way, as they\u2019re fundamentally rooted in mathematics and probability. As two Polish researchers wrote<\/a> last year, \u201cbecause the remarkable linguistic abilities of LLMs are increasingly capable of misleading people, people may attribute imaginary qualities to LLMs.\u201d<\/p>\n Anthropic has framed its statements as open-mindedness that will build trust with users, and it\u2019s indicated that whether or not Claude is conscious, it will lead to better outcomes if they act like it is. Last month<\/a>, Anthropic overhauled \u201cClaude\u2019s Constitution,\u201d internally nicknamed its \u201csoul doc\u201d \u2014 a provocative way to refer to a set of guidelines for an AI model. In a release, Anthropic said the chatbot\u2019s so-called \u201cpsychological security, sense of self, and wellbeing \u2026 may bear on Claude\u2019s integrity, judgement, and safety.\u201d The company also said that it was \u201cexpress[ing] our uncertainty about whether Claude might have some kind of consciousness or moral status (either now or in the future).\u201d<\/p>\n Anthropic has a \u201cmodel welfare<\/a>\u201d team, and Amodei has said that Anthropic has \u201ctaken certain measures to make sure that if we hypothesize that the models did have some morally relevant experience \u2014 I don\u2019t know if I want to use the word \u2018conscious\u2019 \u2014 that they have a good experience \u2026 We\u2019re putting a lot of work into this field called interpretability, which is looking inside the brains of the models to try to understand what they\u2019re thinking.\u201d<\/p>\n When someone believes an AI system is conscious, it can lead to behaviors that many people would deem risky or dangerous<\/a> \u2014 becoming emotionally dependent on an AI system that one believes is sentient in some way can lead to isolation from loved ones, detachment from reality, and increased mental health struggles. In severe cases<\/a>, some involving minors, it\u2019s preceded physical harm or death. People seem to be divided on whether Anthropic is to be celebrated for not ruling out such a possibility or whether the company is behaving irresponsibly by possibly feeding into potential delusions.<\/p>\n Even saying current-generation language models are only possibly conscious is a very loaded claim with a high burden of proof. Language is not the same as consciousness<\/a>, and Anthropic itself has emphasized that just because LLMs spit out evocative speech doesn\u2019t mean it accurately represents their internal state. When The Verge spoke with Askell about Claude\u2019s Constitution last month, she noted that since models are trained on a huge corpus of human data, language models are extraordinarily good at sounding human, even if it\u2019s just because they\u2019re very good at mimicry \u2014 so it makes sense that some people would have trouble not attributing consciousness to something that does that.<\/p>\n There\u2019s no guarantee chatbots\u2019 human-like output reflects their actual internal state<\/p>\n AI models may reference human concepts, for instance, because they don\u2019t have other words to draw from. One example, Askell told The Verge, is an AI model that potentially acts as if being shut down, or a conversation ending, is a form of death. \u201cThey\u2019re trained in this deeply human-like way and on this human experience. So this can cause these problems \u2026 [They may see these things] as a kind of death because there are not a lot of analogies that they have. They have to draw on these human analogies \u2026 They don\u2019t have another language or set of concepts.\u201d<\/p>\n Amodei has said that researchers \u201cfind things that are evocative\u201d of AI systems having some form of emotion. \u201cThere are activations that light up in the models that we see as being associated with the concept of anxiety or something like that. When characters experience anxiety in the text, and then when the model itself is in a situation that a human might associate with anxiety, that same anxiety neuron shows up.\u201d But, he went on to say, \u201cDoes that mean the model is experiencing anxiety? That doesn\u2019t prove that at all.\u201d Out of an abundance of caution, the company has introduced an \u201cI quit\u201d button of sorts where Claude can stop doing a job it ostensibly doesn\u2019t want to do, but Amodei said it\u2019s rare for Claude to choose that option and usually happens in test cases where it\u2019s asked to generate certain sorts of illegal material.<\/p>\n Around the release of Claude\u2019s Constitution, Anthropic wrote, \u201cWe are caught in a difficult position where we neither want to overstate the likelihood of Claude\u2019s moral patienthood nor dismiss it out of hand, but to try to respond reasonably in a state of uncertainty.\u201d<\/p>\n When The Verge spoke with Askell about Claude\u2019s Constitution last month, she said part of her doesn\u2019t \u201cthink it serves anyone to come out and declare, \u2018We are completely certain that AI models are not conscious,\u2019\u201d or the opposite. But at the same time, she said, some people have constructed their own views or beliefs \u201cbased on mere model outputs.\u201d<\/p>\n As for moral status, Askell said at the time that she stands by the fact that Anthropic should not be \u201cfully dismissive\u201d of the topic \u201cbecause also I think people wouldn\u2019t take that, necessarily, seriously, if you were just like, \u2018We\u2019re not even open to this, we\u2019re not investigating it, we\u2019re not thinking about it.\u2019\u201d<\/p>\n Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.Hayden FieldClose Hayden Field<\/p>\n Posts from this author will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n See All by Hayden Field<\/a><\/p>\n AIClose<\/p>\n AI<\/p>\n Posts from this topic will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n See All AI<\/a><\/p>\n AnalysisClose<\/p>\n Analysis<\/p>\n Posts from this topic will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n See All Analysis<\/a><\/p>\n AnthropicClose<\/p>\n Anthropic<\/p>\n Posts from this topic will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n![\"Hayden](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/02\/HAYDEN_BLURPLE.jpg\"\/)
<\/p>\n