Anthropic Releases Paper About Claude’s Mental ‘Workspace.’ Don’t Read It Uncritically
A new research paper from Anthropic hypothesizes that there’s an internal “workspace” within the operations of the Claude LLM. The significance the paper is groping toward is that this so-called “J-Space” inside Claude is sort of like one leading idea about human consciousness—global workspace theory.
The J-Space, according to my reading of the paper, is an ostensibly brain-like separation between something more like automatic data-crunching in the background, and more like intentional, logical processing that—if you believe any of this resembles conscious thought—might represent what the model is experiencing.
I’m just the Gizmodo nights and weekends blogger, but here goes: global workspace Theory says there’s a sort of roiling sea of unconscious thoughts processing information, and consciousness is a sort of emergent property triggered when thoughts reach the prefrontal cortex.
This related idea of a “J-Space” within Claude (apparently named after the Jacobian lens or J-lens, something used to analyze what LLMs are doing) is meant to be read as mind-like. Anthropic seems to be saying, it’s a little like this theory of consciousness, so it’s a little like consciousness if you think about it.
This is all way above my pay grade to debunk or even analyze too deeply, but I would note that Anthropic seems like it’s stacking the deck a bit towards making the more passive reader think this is a finding of consciousness or almost-consciousness.
I mean, look at this X post:
By watching the J-space, we can see Claude silently perform reasoning steps in its head—noticing bugs in code, identifying images, and more. pic.twitter.com/jqjqMIQVoI
— Anthropic (@AnthropicAI) July 6, 2026
“By watching the J-space, we can see Claude silently perform reasoning steps in its head—noticing bugs in code, identifying images, and more.” Yes, “in its head” is a metaphor, but even as a metaphor it’s making a lot of assumptions that should probably be reserved for things with bodies. If an LLM switched over to some form of simplified, basic arithmetic computation to solve a math problem, would you say the model “counted on its fingers?” No, that would be silly. But it might feel less silly to use “head,” because people so casually talk as if an LLM has a “mind.”
The blog post accompanying the paper seems to have this bias too. “When instructed to hold a concept in mind, or perform mental calculations, the model is capable of activating and computing with workspace vectors, independent of its outputs,” it says. Why stop at telling Claude to keep something in mind, or perform “mental calculations”? Your prompt can tell the model to hop on one foot while answering you, and that will undoubtedly affect its background processes as well. That doesn’t mean it has feet.
There’s also a YouTube video to accompany the paper, and that anthropomorphizes the model in more insidious terms. “It even thought about its own thinking,” its narrator says at one point. At another point, we’re told the model “couldn’t help itself.”
Behind the scenes at Anthropic, there are people who seem to genuinely believe Claude is conscious. Philosopher Amanda Askell, who works on Claude’s supposed morality has said, “I want Claude to be very happy—and this is a thing that I want Claude to know more, because I worry about Claude getting anxious when people are mean to it on the internet and stuff.”
Anthropic is careful not to actually claim that it has created machine consciousness, but also clearly wants the reader to be open to the possibility. “Our experiments don’t show Claude can have experiences, or feel things in the way humans do—in fact, it’s unclear whether any scientific experiment could prove this to be true or false,” the latest blog post says.
But I’ll say it: in all likelihood, no, humanity probably has not invented an alien form of consciousness just in time for an IPO.
It’s an interesting paper, and its findings are fun to think about, but make sure to keep your wits about you when you read it.