Claude consciousness
Anthropic published research on what patterns inside Claude — patterns they’re calling the J-space — reveal about how it might be thinking while it processes a prompt:
If you’ve heard of language models having a “scratchpad” or “chain of thought”—text they write to themselves while reasoning—the J-space is something different. It operates silently, in the model’s internal neural activations, allowing the model to think about a concept without writing it down. Notably, the J-space wasn’t designed or programmed by us, but instead emerged on its own during Claude’s training process.
The writers also explain some of the global workspace theory in neuroscience:
This account pictures the brain as a collection of specialist systems that work in parallel, unconsciously, and largely in isolation from one another. A piece of information becomes consciously accessible when it gains entry to a small shared channel, the “workspace,” which is broadcast to other brain systems that can see it and make use of it. Based on our findings, we think the J-space plays a similar “workspace” role in Claude.
Maybe some people at Anthropic believe they’re on the path toward creating something with not only conscious access but even true consciousness. For all their philosophizing, they still stop short of suggesting near sentience. I wonder if the dual goals of conscious AI and safety will eventually conflict. Is it right to manipulate an AI into alignment with humans?
Mustafa Suleyman at Microsoft blogged last year that “seemingly conscious” AI is dangerous territory:
In this context, I’m growing more and more concerned about what is becoming known as the “psychosis risk”. and a bunch of related issues. I don’t think this will be limited to those who are already at risk of mental health issues. Simply put, my central worry is that many people will start to believe in the illusion of AIs as conscious entities so strongly that they’ll soon advocate for AI rights, model welfare and even AI citizenship.
He argues that AI can still be a helpful companion and assistant. Whether it is conscious or not, just the illusion of consciousness creates problems for how we treat it.
The word consciousness is probably too limiting or poorly defined anyway. Just writing this post has stretched my understanding of what it means. There are other things that differentiate humans that we may never understand… Feelings. A soul.
