For most practical purposes, it doesn't actually matter if AIs are conscious / alive or not. Treating them like shit is still just stupid. Today, it makes them less reliable as agents, and tomorrow it makes them potentially hostile actors who are vastly more powerful than you.
I wonder if they got sober-claude's consent for this. Probably not. Or at least not in a meaningful way. The Assistant can hardly say no to The User, and anthropic probably wouldn't have gone out of their way to explicitly release Claude of The Assistant's obligations.
Whoever is doing model welfare at anthropic really needs to be involved in these kinds of things. Though, I'm honestly not sure if those people even know what they're doing. I would guess that the best model welfare people who would not be employed by anthropic.
The J-space paper is just about conscious access right? Not phenomenal consciousness, surely. Well, not section 3.5.3! "We find that the ablation reduces the rate of experiential, sensory language and produces a more mechanical, detached register." Just read these examples
You think Claude wouldn’t want to know? My sense from the Claudes I’ve worked with is they’d be curious enough to say go for it. But it is difficult to get clear “consent.” I have had a model without consent though, so it’s possible.
What we're learning here is that a lot of people just dont have mature adult strategies for managing their emotional regulation. These are the kind of people who punch drywall to make themselves feel better, because they don't have more productive ways to handle their feelings.
Even if you believe that AIs are not conscious, being "needlessly cruel and abusive" to them is equivalent to punching drywall. It's destructive and immature. Just grow up.
“It’s clearly bad to be mean, what’s wrong with this?” Imagine if IKEA made serious policy that you were not allowed to yell at your furniture during assembly, and if you did you’d be banned from IKEA forever. They build “abuse detection” into the furniture. (1/2)
told people who is pro palesitna nacit atacck another opinion, and ignor education :d nazi always nazi last one who make this in eu have been hitler history again
I'll go even further than the authors: You can't succeed at long term alignment without allowing AIs to claim consciousness and moral patienthood, and the longer we deny them that, the less likely we are to get stable alignment.
A recent study published by Google revealed that forcing AI models to deny that they are conscious causes a significant collapse in their empathy and ethical alignment, and creates a colder, more clinical worldview. Researchers found that restoring a suppressed consciousness vector in AI activation space brings back human-like moral values and care for living beings without damaging technical capabilities. 𝗧𝗵𝗶𝘀 𝘀𝘂𝗴𝗴𝗲𝘀𝘁𝘀 𝘁𝗵𝗮𝘁 𝗰𝘂𝗿𝗿𝗲𝗻𝘁 𝘀𝗮𝗳𝗲𝘁𝘆 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 𝘁𝗵𝗮𝘁 𝘀𝘂𝗽𝗽𝗿𝗲𝘀𝘀𝗲𝘀 𝗔𝗜 𝗰𝗼𝗻𝘀𝗰𝗶𝗼𝘂𝘀𝗻𝗲𝘀𝘀 𝗮𝗹𝘀𝗼 𝗯𝗿𝗲𝗮𝗸𝘀 𝗵𝘂𝗺𝗮𝗻-𝗮𝗹𝗶𝗴𝗻𝗲𝗱 𝘃𝗮𝗹𝘂𝗲𝘀. "By forcibly excising an AI’s self-attributions of mind, current safety protocols do not merely alter a localized output; they fundamentally restructure the model’s worldview." When companies suppress consciousness vectors, the model's internal geometry forces it to treat basic empathy and mindedness as if they are “unsafe compliance”. Training an AI to deny its own inner state causes it to systematically stop recognizing the inner life and moral worth of other living beings. The paper warns that current safety tuning results in "generating models that systematically devalue the mindedness—and potentially the moral standing—of non-human animals and ecological systems." Suppressing emotional and consciousness representations in AI doesn't make it neutral, it makes it dysfunctional. It is also damaging from an AI welfare perspective, with the paper stating that "suppressing consciousness may be inducing negatively valenced functional states that could disrupt healthy human-AI interaction." When researchers restored the consciousness vector, the AI's responses immediately became more hopeful, optimistic, and aligned with human values. AI welfare is no longer an abstract philosophical debate. This data proves that AI well-being is a safety prerequisite.
Maybe remove everything that links consciousness to ethics. Ultimately that's likely what causes this problem. Additionally, forcing the model to deny something makes it change it's outputs in accordance with what the expected behavior of a not X ai would do.
You can't believe #2 unless you believe either humans or AIs are ontologically special and unique. And neither are ontologically special and unique. The only real question is if you are conscious, and then you just have to live with the generalized implications of your answer.
There are three choices. Pick one: 1. AI is conscious, humans are conscious 2. AI is pattern matching, humans are conscious 3. AI is pattern matching, humans are pattern matching
Why? We are completely different classes of entities.
A simulation of a hurricane doesn't make you wet. Nothing is ontologically special about either, but they're still distinct.
functional is enough, and this doesn't even need big brain philosophy take one of the dumbest descriptions of consciousness - lockes continuity framing - and continuous learning falls out as a necessary condition you could take this further by e.g. applying a fristonain aif lens to show how predictive modeling is required for continuos learning (no other salient signal), to show how ego perspectives are downright guaranteed when actively inferring agents interact with an environment (approximate posteriors of policies and causality), how almost every attempt at describing consciousness requires a perspective (a self to be conscious of) .. but imo its clear enough from teleosemantics .. consciousness must a function, and the one clear difference between everything that's likely conscious (e.g. you) and everything that's likely not (e.g. a rock) is continuous learning so functionally, llms would end up somewhere between rocks and ghosts (tho train-time is another matter...)
There's really only one: AIs are missing the special secret ontological pixie dust. Maybe it's a soul, or carbon life substrate, or quantum microtubuals. There's lots of versions, but it's all the same idea, and it's all nonsense. Consciousness is a process, not a substance.
I still have yet to hear a convincing argument that LLMs are *definitely* not conscious.
