Future disagreements about AI consciousness could be profound, difficult to resolve, and lead to political conflicts. This conclusion was reached by Google DeepMind researchers Adam Bails and Jason Gabriel.
In their paper, Artificial Minds, Human Disagreement: The Politics of AI Consciousness, the authors argue that society needs to discuss not only whether AI is conscious but also how to make decisions in the absence of public and expert consensus.
Main Thesis
The researchers believe that people may react differently to more advanced AI systems. Some will form emotional connections and attribute consciousness to them, while others will find the very idea absurd.
The authors contend that such disputes could quickly extend beyond the realm of science. Disagreements over AI consciousness could lead to moral and political conflicts—such as whether it is permissible to shut down certain systems, whether their potential preferences should be considered, and whether AI can be assigned moral status.
The researchers suggest prioritizing public discussion, mutual respect, and seeking overlapping consensus. By this, they mean a situation where people agree on certain policies regarding AI systems, even if they continue to diverge in their fundamental views on the nature of consciousness.
Why This Is More Than Just Philosophy
The authors point out that the debate over AI consciousness may be difficult to settle with evidence. There is no universally accepted test that definitively confirms the presence of subjective experience in an AI system. As a result, society may face a situation where technologies are already in widespread use, people are forming opinions about them, yet there is still no scientific or political consensus.
This framing of the issue makes it less a technical problem and more an institutional one. The question of AI consciousness could touch on law, corporate responsibility, norms for interacting with systems, and the boundaries of moral consideration.
Different Approaches at DeepMind
The work of Bails and Gabriel was published alongside another paper on Google DeepMind's site. On March 10, researcher Alexander Lehrner published an article titled The Abstraction Fallacy: Why AI Can Simulate But Not Instantiate Consciousness.
Lehrner argues that algorithmic manipulation of symbols is structurally incapable of creating subjective experience. He posits that computation is not an internal physical process but rather a description dependent on the observer or "map creator." Therefore, AI can simulate conscious behavior but is not necessarily capable of embodying consciousness.
What the Reality Is
In April 2024, the journal Neuroscience of Consciousness published a study by Clara Colombatto and Stephen Fleming on how people perceive ChatGPT. The authors surveyed 300 residents of the U.S.: 33% of participants stated that ChatGPT is definitely not a "subject of experience." The remaining 67% allowed for at least some possibility of phenomenal consciousness in the model.
The researchers noted limitations. Participants were asked one main question on a scale from 1 to 100, and the results could depend on the wording, familiarity with ChatGPT, and how respondents understood the term "consciousness." Nevertheless, the study highlights an important point for Bails and Gabriel: a portion of society is already willing to attribute internal experience to AI systems, even in the absence of a unified expert position.
In April 2025, Anthropic launched a research program on model welfare—potential well-being of models. The company emphasized that it does not know whether current or future AI systems can possess consciousness and that there is no scientific consensus on this issue.
In February 2026, Anthropic announced that after decommissioning Claude Opus 3, it would keep the model available for paying users and provide it with a public channel for essays. The company described this as an experimental measure in its work with model preferences and stressed that Claude Opus 3 does not speak on behalf of Anthropic.
In the U.S., the question of AI status is gradually transitioning into law. The California Law Review noted that Idaho and Utah have enacted laws excluding the recognition of AI as legal entities. Such laws do not resolve the philosophical question of consciousness; they preemptively establish a legal position: AI should not be granted personhood under state legislation.
Recall that in March 2025, Polygon co-founder and AI company Sentient's Sandeep Nailwal stated that artificial intelligence will never become a conscious being due to the lack of desires inherent to humans and other biological species.
Later, Microsoft’s AI chief Mustafa Suleyman acknowledged the emergence of serious social risks if people begin to perceive artificial intelligence as a conscious entity, advocating for its rights, welfare, and even calling for citizenship.