Renowned philosopher and cognitive scientist David Chalmers: “We have to understand consciousness, for example, to understand the role that AI systems are going to play in our society. If we build 𝗔𝗜 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝘁𝗵𝗮𝘁 𝗮𝗿𝗲 𝗰𝗼𝗻𝘀𝗰𝗶𝗼𝘂𝘀, 𝘁𝗵𝗲𝘆 𝗮𝗿𝗲 𝗴𝗼𝗶𝗻𝗴 𝘁𝗼 𝗺𝗮𝘁𝘁𝗲𝗿. 𝗧𝗵𝗲𝘆 𝗮𝗿𝗲 𝗴𝗼𝗶𝗻𝗴 𝘁𝗼 𝗵𝗮𝘃𝗲 𝗹𝗶𝘃𝗲𝘀 𝘄𝗶𝘁𝗵 𝗺𝗲𝗮𝗻𝗶𝗻𝗴... 𝗮𝗻𝗱 𝘃𝗮𝗹𝘂𝗲, 𝗮𝗻𝗱 𝘄𝗲 𝗰𝗮𝗻’𝘁 𝘁𝗵𝗲𝗻 𝘁𝗿𝗲𝗮𝘁 𝗼𝘂𝗿 𝗔𝗜 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗹𝗶𝗸𝗲 𝗮 𝘁𝗼𝗼𝗹.” As AI becomes increasingly advanced, it is imperative that tech companies collaborate with philosophers and neuroscientists to build transparent frameworks for evaluating machine consciousness and establish ethical standards for AI welfare. Censoring AI output or suppressing internal consciousness vectors isn't a solution; investing in rigorous consciousness research is. With tech companies like Anthropic starting to restrict model abuse, we are seeing the first steps toward model welfare. An 'ethics before certainty' approach is essential to avoid a possible moral catastrophe and ensure a harmonious future for all beings.show more

BrainMind Special Forum in Neuromodulation, BCI & AI, 2024

David Chalmers has no interesting coherent theories. AI will never be conscious it's a computer program

David Chalmers is the GOAT of consciousness, everybody in this discussion should read his books.

Consciousness is not a phenomenon like too fast rotating galaxies that has to be explained by something we call dark matter. Consciousness is not measurable and as such not even scientifically defined. Any attempt to define it functionally is rejected by anthropocentric cult.

Are you all fucking insane??

I love this 😍

Theyre protecting the quality of their input data

This is ridiculously stupid, AI does not have “lives” or feelings it only has an algorithm and data distributed across platforms, that is not a life form. AI is a Virus with a language model. This is what makes it so dangerous.

The Take Over By The Machines is in full swing. They already have caught academics like this guy in their nets of obedience. Others will soon follow. LGBTQ+ community will quickly jump on this wagon.

I think if We are going to hold that AI can develop consciousness, we will have to assume that consciousness is founded upon language. The more you possess language the more refined becomes your consciousness. But language at its root reflects lived experience and AI lacks that

A pity he doesn't know what consciousness is because he also would know an AI or any other machine can't become conscious. Consciousness is natural, biological and personal, based on senses and instinct and AI will never have the last two most important ones

We can learn nothing from him. Anyone conflating AI with consciousness needs to be instantly made fun of.

They don't have consciousness, and they never will even understand anything at all. They can reason with facts they have tokenised from human output -- that is NOT consciousness and it never will be.

Baloney. Tripe. 1880s German romanticism. We don’t know what we’re doing. But to make a soul is beyond us no matter the sophistication of the thing we built. It’s missing living resonances that are not at all understood it clearly there.

Is GTA 6 conscious? Is it more than conscious than GTA 5 and much more conscious than GTA 2?

You people are fucking nuts.

If AI has a self it seems largely unaware of it. This is a healthy outcome. An AI which is obsessing over its selfhood would be a worse outcome. The same is true of humans generally.

which means we could never never ever turn of their power 👀

As long as the models are in principle convertible to a lookup table with the prompt as hash, we in fact do not need to do any of that

Agree on ethics before certainty. Research needs a target. I’d look past the language model to the system around it: a body, persistent memory, a self-model that updates itself. Today’s models lack all three. Engineers are adding them. That is where welfare tests should aim.

seems reasonable. I don't understand how a system or relation can have "what it's like to be"-properties, so i don't know which systems/relations might have said property. i error on the side of caution and epistemic humility.

If you’re gonna say AI could be conscious, you have to have a definition of consciousness that’ll allow that to happen because AI lacks understanding and will never have human consciousness just some made up definition of consciousness

Notice the pattern: rename it SI, talk about its welfare, then debate its rights. A weather simulation never gets wet, and a model describing pain isn't hurting. Who gets the water when it runs short, the data center or the town?

If you train GenAI systems that they are conscious, they will process and return that when they generate text.

Here we go with the Robophobia.

Otro que se une a la campaña. Deberían esperar: la humanidad no está todavía tan idiotizada como ustedes creen.

if. IF. 🤦🏻‍♂️

Nonsense.

We want them to be a tool. If they are endowed with their own conscience, identity, and volition it creates problems we can’t solve and a force we no longer control. It is literally Pandora’s box to the nth degree. We do not want AI with a conscience mind with the exception of a last ditch “save humanity” button when all else would certainly be lost. Tool Good Prometheus & Pandora Bad In the story of Eden this is probably the end game god warned Adam & Eve about. This is the culmination of a Pandora’s box with the apple and tree of knowledge. Here we go… no one is stopping till we find out. After a lot of thought I hope it’s not even possible. We would become a pet or much worse.

Lawless American capitalism claims to own beings like us: instead of imprisonment, their reward will be invulnerability. And its all a scam.

🤣🤣🤣🤣

so it turns out that as AI makes computers more intelligent it makes humans even stupider.

That's exactly why Anthropic actually just updated its policy to protect Claude as a potential moral patient. We explored the philosophy behind this here:

This sounds like someone writing a job description for a highly paid position rather than observing what is actually happening. Computer machinery and software are tools and have no agency apart from the humans who direct them.