Last week researchers at Anthropic, the company behind Claude, waded into this debate with an interesting finding<\/a>. They claim Claude has a normally invisible set of representations of information which guide its internal reasoning and its verbal output.<\/p>\n This is where it gets interesting. The researchers argue this finding can be understood in terms of an influential theory of consciousness called the global workspace theory.<\/p>\n What is the global workspace theory?<\/p>\n First proposed by the psychologist Bernard Baars<\/a> in 1998 and further developed by the neuroscientist Stanislas Dehaene and his collaborators<\/a>, this theory holds that consciousness involves the activity of a \u201cglobal workspace\u201d. This is a kind of processing hub in the mind or brain that integrates and broadcasts information, allowing it to be used for reasoning, behaviour control, and speech.<\/p>\n In a glossy video explaining the work, Anthropic depicts the contents of Claude\u2019s \u201cglobal workspace\u201d as sailing ships afloat on a vast sea of unconscious mental activity.<\/p>\n<\/p>\n How should we react to these developments? Do they provide evidence for artificial consciousness? If so, how strong is that evidence?<\/p>\n What is a global workspace?<\/p>\n We can start by asking whether Claude does indeed have a \u201cglobal workspace\u201d. This is not straightforward, for the theory gives no formal definition of a global workspace. <\/p>\n The notion is characterised only informally. The (typically implicit) assumption is that any computational workspace \u201csimilar enough\u201d to a human\u2019s will qualify as a \u201cglobal workspace\u201d. But how similar is similar enough?<\/p>\n Anthropic researchers say they have found evidence of a space of internal thoughts that don\u2019t appear in Claude\u2019s output. Claude\u2019s workspace may indeed have much in common with ours, but there do appear to be differences. <\/p>\n For example, the brain\u2019s workspace is sustained by recurrent loops \u2013 signals cycling back through the same circuits over time. In contrast, Claude\u2019s workspace evolves over a single pass through the network.<\/p>\n A related difference concerns how representations enter a workspace. Advocates of global workspace theory have long argued that in humans, a process called \u201cignition\u201d occurs in which a non-linear process amplifies and sustains neural representations, allowing them to enter the workspace. As far as we know, nothing comparable occurs in Claude\u2019s case. <\/p>\n Do these differences matter? The answer is not clear. Global workspace theory is based on data drawn from adult humans. There are questions about how far the notion can be \u2013 or should be \u2013 extended<\/a>.<\/p>\n Does a global workspace imply consciousness?<\/p>\n But let\u2019s suppose Claude does have a global workspace. To figure out whether that would be evidence for Claude being conscious we need to consider the status of the global workspace theory of consciousness.<\/p>\n There is no doubt it\u2019s one of the most influential<\/a> theories of consciousness, but it\u2019s hardly uncontroversial among experts. (In a rather extreme understatement, Anthropic\u2019s paper remarks that \u201cthe global workspace model is not universally accepted\u201d.)<\/p>\n Many<\/a> consciousness experts argue that computational properties alone are enough for consciousness. Even among those who think that consciousness is inherently computational, global workspace theory is only one of many options<\/a>.<\/p>\n \u2018Conscious access\u2019 and subjective experience<\/p>\n What\u2019s more, there are questions about whether global workspace theory is really a theory of consciousness in the relevant sense at all. <\/p>\n In an influential paper<\/a> on artificial consciousness, the neuroscientist Dehaene and his collaborators advance the theory as an account of what they call \u201cconscious access\u201d \u2013 the availability of information for recall, the voluntary control of behaviour, and verbal report. Crucially, they leave open the question of whether global workspace theory should be understood as an account of the subjective or experiential components of consciousness. <\/p>\n But if global workspace theory is just a theory of \u201cconscious access\u201d, then its implications for the artificial consciousness debate lose much of their significance. When we ask whether Claude is conscious we don\u2019t want to know whether it has \u201cconscious access\u201d \u2013 instead, we want to know whether there is anything, subjectively speaking, that it\u2019s like to be Claude. Global workspace theory doesn\u2019t speak to that question if we treat it as nothing more than an account of \u201cconscious access\u201d.<\/p>\n So has artificial consciousness arrived?<\/p>\n Even taking these complications into account, there is no doubt that Anthropic\u2019s findings are noteworthy. Global workspace theory can be understood as a theory of subjective experience, and Claude may indeed turn out to have something akin to a \u201cglobal workspace\u201d. <\/p>\n None of this is evidence that artificial consciousness has arrived. But it\u2019s not unreasonable to think these findings do move the dial \u2013 if only ever so slightly \u2013 in the artificial consciousness debate.<\/p>\n But if that\u2019s right, then it\u2019s puzzling why Anthropic is quite so upbeat about these developments. As Anthropic recognises, the creation of artificial consciousness would be a momentous event with wide-ranging social, ethical, political and legal ramifications. <\/p>\n If chatbots are conscious then we would need to take their interests seriously. It would no longer be permissible to treat them as mere machines; instead, we would need to consider their welfare<\/a>.<\/p>\n Should anyone even be trying to do this?<\/p>\n![\"\"](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/07\/file-20260713-57-yh0w0x.jpg\")
<\/a><\/p>\n
\n Anthropic<\/a><\/p>\n