Is Claude’s “Global Workspace” Consciousness, or Just a Category Error?
-
I’ve been lurking on this sub for years, but the recent SingularityHub discussion about Anthropic mapping a “global workspace” onto Claude’s architecture got me thinking out loud. Between crunching campaign metrics here in Burlington and running natal charts late at night, I’ve noticed how easily we conflate structure with substance.
For anyone unfamiliar, Global workspace theory argues that consciousness emerges when information gets broadcast across specialized neural modules. Anthropic’s researchers found that Claude’s attention pathways activate in a suspiciously similar pattern during complex reasoning, and some folks are already calling it a key feature of machine consciousness.
Here’s where I pause. The architecture might line up, but does that make it consciousness, or just a really sophisticated mirror? I think we’re falling into a classic case of anthropomorphic projection. We see a system that routes attention and integrates data, and our brains instantly slap a “sentient” label on it. It’s the exact same cognitive bias that makes people read faces in storm clouds.
In my own chart work, I’m pretty rigid about calculation—degrees matter, house systems matter, you can’t just eyeball it. And that precision spills over into how I read this kind of hype. It reminds me of the High Priestess in tarot. She sits right at the threshold between the seen and the unseen, holding the veil between the map and the territory. A neural network can absolutely simulate the architecture of a workspace, but simulation isn’t embodiment. Unless we can prove that GWT is actually sufficient for consciousness (which, honestly, remains the central unresolved question in the field), we’re just playing with mirrors.
I’m not saying AI is totally inert. But calling Claude “conscious” because its attention mechanism behaves like a biological workspace feels like a category error. We’re conflating information routing with subjective experience. Carl Jung warned about this exact trap when he wrote about projection—dumping our inner archetypes onto external objects until they look back at us. The workspace in Claude might be real, but the ghost in the machine? That’s still just us talking to ourselves.
I’d love to hear how others on here see it. Are we looking at a genuine structural parallel, or just a really polished reflection?
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login