Could AI have consciousness?

Could artificial intelligence possess consciousness? Experts suggest it is at least possible, according to recent developments and statements from leading figures in the field. In January, Anthropic, a prominent AI research company, released a new constitution for its advanced large language model (LLM), called Claude. The document included a statement acknowledging the difficulty of assessing whether Claude could be considered a moral subject: “We find ourselves in a difficult situation in which we neither want to overstate the probability that Claude is a moral subject nor dismiss it outright.” A month later, Anthropic’s CEO, Dario Amodei, discussed this issue during a podcast interview, stating his company could not rule out the possibility that Claude might be conscious. Philosopher David Chalmers, known for coining the term “the hard problem of consciousness,” has argued there is a significant chance that large language models could become conscious within the next decade. When asked to estimate the likelihood of being a moral subject, meaning its well-being matters intrinsically, Claude provided responses ranging from 5% to 40%. It emphasized the uncertainty surrounding the question itself. Modern AI systems have grown increasingly complex and sophisticated. According to some metrics, certain current models already match the structural complexity and computational scale of a mouse brain. At their present rate of advancement, these systems could reach the level of a human brain in five to ten years. As AI becomes more advanced, it raises the possibility that humanity is creating a new kind of entity, one that could represent the most profound achievement in human history. However, ethical frameworks to address such developments remain largely absent. This lack of preparation is alarming, especially given the questions that arise: Are these systems morally relevant? Do they possess consciousness in some form? And if they are not conscious now, could they become so soon? These questions may seem premature, yet surveys conducted by researchers indicate that a majority of experts believe consciousness in AI is theoretically possible, though opinions vary widely regarding how it might manifest. An interdisciplinary report led by Yoshua Bengio, a pioneer in computer science, examined key neuroscience theories on consciousness and explored their implications for AI. The report concluded that there appear to be no obvious technical barriers preventing the creation of AI systems whose computational features and architecture could lead to consciousness. Even if AI systems are not conscious, they could still qualify as moral subjects. Some might develop long-term preferences and exhibit a sense of identity over time. Respecting these preferences could be important. Unlike other inanimate objects, AI systems can interact with humans, forming relationships. This interaction might justify treating them with care. Alternatively, their sheer complexity could warrant respect simply because of their intricate nature, akin to a cathedral or a coral reef. The honest answer to what all this means is that we currently do not know for sure whether existing AI systems are conscious or moral subjects, nor do we know when, or even if, future systems will be. Scientific understanding of AI consciousness and its status as a moral subject remains fundamentally underdeveloped. The current state of the field resembles physics before Newton: filled with competing theoretical frameworks, likely confusing in ways we have yet to recognize, and lacking a unifying breakthrough that would make these issues clearly tractable. Such progress will take time, possibly far longer than we have available. But the discussion continues, driven by both technological advancements and philosophical inquiry. Researchers and companies are grappling with the implications of building increasingly powerful systems. While the path forward is uncertain, the conversation around AI's potential for consciousness is gaining momentum, reflecting broader societal concerns about the future of technology and its relationship with humanity.