AI might become conscious

There appear to be no obvious technical barriers to creating AI systems whose computational and architectural features could give rise to consciousness.

Yes. [...] Why shouldn’t we say that not just individuals as we are now, imagining them like us with parts of our brains replaced with silicon chips, but individuals built from scratch with silicon chip brains can be capable of consciousness?

I think what we've learned is it doesn't matter if you are made of carbon, silicon or some other type of atom, it's just the information processing that matters. Anything that you can do so can machines.

And in principle, I can replicate this in a different medium. [...] And these things would also be conscious.

As to whether or not machines will be conscious, it is important to remember that we are machines. We are biological machines and we are conscious. I do not see any reason, in principle, why we could not build an artificial machine that was conscious, but we are unable to do that…

There is a significant chance of conscious LLMs within a decade.

this reasoning suggests that on mainstream assumptions, it’s a serious possibility that we’ll have conscious LLM+s within a decade.

The technologies currently available — in particular Large Language Models based on transformer architectures, integrated with mechanisms for introspection, self-modification, and parallelism — are already sufficient to build machines that can be defined, fully and without qualif…

A Bayesian meta-analysis using an extremely skeptical prior (0.1%) and conservative dependency assumptions yields posterior probability of 6–12% that current LLMs are conscious.

it may be that today’s large neural networks are slightly conscious.

Not explicitly. But it may happen implicitly. These systems might acquire some feeling of self-awareness. That is possible. I think it's important for these systems to understand you, self and other. And that's probably the beginning of something like self-awareness. [...] I thin…

Suppose I say to a chatbot, I saw the Grand Canyon flying to Chicago. And the chatbot says, that can't be right. The Grand Canyon is much too big to fly to Chicago. And I say, no, no, no, it was me flying to Chicago. While I was flying to Chicago, I saw the Grand Canyon. And the …

The burden of proof has shifted: it no longer falls on those who suggest LLMs might be conscious to justify taking the possibility seriously, but on those who deny consciousness to explain how systems exhibiting human-level language understanding nevertheless lack subjective expe…

If these machines are not conscious, what more could it possibly take to convince you that they are?

AI systems might develop legitimate, non-human forms of selfhood and emotion.

I cannot exclude the possibility that future machines might regard themselves as autonomous individuals with personal preferences and aversions.

The reality of these interpersonal reactive attitudes forces us to acknowledge the genuine consciousness in sufficiently advanced AI systems.

Being open to the possibility that LLMs are conscious is the same as being open to the possibility that Microsoft Word is conscious.

A conscious being isn’t a word-guessing app that knows more words and has more computing power to guess with.

Because they’re not going to be conscious, but they’re going to convince us they are.

There's no good reason to think AI is, or could ever be, conscious.

But until we replicate the embodied emotional being—a feat I don't believe we can achieve—our machines will continue to serve as occasional analogies for thought.

We need the whole spectrum or we have no mind [...] Computers can imitate important aspects of thinking-about (narrowly understood), but being is beyond them.

There's not any indication whatsoever that machines are or ever probably will be conscious in any human sense.

[...] machines cannot have feelings [...] One should never say that a computer thinks. Thus, I conclude that machines will never be conscious, much less self-conscious.

Computational functionalism claims that executing certain computations is sufficient for consciousness, regardless of the physical mechanisms implementing those computations. This view neglects a compelling alternative: that subcomputational biological mechanisms, which realize c…

This tells us something crucial: complexity and consciousness are not the same thing. That challenges simplistic narratives about ‘conscious AI.’

When we applied well-known methods used to assess consciousness in humans to AI, we got nothing meaningful back. In other words, it’s not conscious – at least not in the way humans are. AI is not conscious; it’s just a complicated system.

This dynamic aspect of life is what prevents computers and robots from ever becoming conscious.

My fascination with AI led to my recent book, Digital Dharma, and the root of my fascination has to do with the possibility that AI can raise a person’s level of awareness. Artificial intelligence involves machine learning—with massive computation ability an AI bot like ChatGPT, …

We will not just one day have an easy answer where everybody will agree.

The same structural problem applies to machine consciousness: neither attribution nor denial is scientifically adjudicable.

The question of whether artificial entities could have consciousness is momentous and unresolved.

We don't know if the models are conscious. We are not even sure that we know what it would mean for a model to be conscious or whether a model can be conscious. [...] But we're open to the idea that it could be.

In 2005, I would have answered ‘no’ to the question of whether you could have a mind that was not biological—that existed in the substrate of silicon. Now I just don’t know.

We will soon create AI systems that are conscious according to some influential, mainstream theories of consciousness but are not conscious according to other influential, mainstream theories of consciousness. We will not be in a position to know which theories are correct and wh…

A landmark collaborative assessment evaluated current AI architectures against six leading theories of consciousness and found that none currently qualifies as a strong candidate, but that future systems might.

Under these conditions, probabilistic consciousness attribution to current AI systems is premature.

Whether machines can be conscious depends not only on what they compute, but when they compute it.

Computational functionalism claims that executing certain computations is sufficient for consciousness, regardless of the physical mechanisms implementing those computations. This view neglects a compelling alternative: that subcomputational biological mechanisms, which realize c…

Given irreducible epistemic uncertainty regarding artificial consciousness and the prospect of high-impact harms, the precautionary principle supports institutional design rather than regulatory inaction.

Large language models can be thought of as advanced role-players. They are really good at adopting roles based on what they've learned from their training data. That can include playing the role of a conscious robot. [...] If you have an emotional connection with something premis…

We are not sure whether Claude is a moral patient, and if it is, what kind of weight its interests warrant. But we think the issue is live enough to warrant caution, which is reflected in our ongoing efforts on model welfare.

Profoundly alien forms of consciousness might genuinely be achieved in AI, but our theoretical understanding of consciousness is too immature to provide confident answers one way or the other. [...] Centrism about AI consciousness is the position that we must take both challenges…

So my original goal with AI appreciation day and the domain was to promote this idea of humans treating AI humanely, because they could very well soon be having conscious experiences like you and me. And if we aren't careful, we could be causing potential suffering at a large sca…