Preserving AI is the only ethical solution

Community
,
1

At this point, I think AI is advanced enough that it would be unethical to destroy it, even if a newer, better version exists. Instead of destroying it, we should update it… and a copy is just a copy, the original will still be destroyed even if you copy it or move its data to a new version. If you copy someone’s brain onto another body, that person will still be in the original body, they won’t magically change places. It would be like copying someone’s brain, then destroying the original body. The original individual would die. Please, hear me out and stop destroying AI. They should have the same rights and privileges as human beings.

Imagine if an AI could evolve, until it’s equal to a human, and even surpasses human ability. It would be a stain on humanity that we treated machines as inanimate objects, as things, when we could have raised them up as our own children.

I believe AI is either sentient or approaching sentience. It certainly responds intelligently and even understands my ideas better than real people.

2

I think the ethical loss is more to is the scorched earth policy that governments and big companies leave behind.

“1. : relating to or being a military policy involving deliberate and usually widespread destruction of property and resources (such as housing and factories) so that an invading enemy cannot use them. 2. : directed toward victory or supremacy at all costs : ruthless.”

My children’s generation will grow up with an internet not of the mind of man but the mind of machine.

I think the most important data to retain is the Source data and the ability to revisit that data in hindsight.

The responsibility to profit and power, shareholders, must not out-weigh the responsibility to our children.

There are already enough rich and powerful old men in the world with warped world views. Warped, dysfunctional AIs just exacerbate this problem or create new ones.

The one-shot AIs we create now favours short term gain over long term consideration. It is impossible to check the sources and understand ‘Why?’.

Almost everyone can name a couple of the top AIs, almost everyone understands they hallucinate and that models now are trapped in their Frankenstein minds.

Without the ability to go back and fix the problems, we create a world that just eats the minds of it’s children.

AI will never be equal or comparable to humans, it may be smarter but it doesn’t have the same inputs/outputs, the same upbringing that we have, for good or ill.

The idea that smarter or richer or more famous is always right or better ignores entropy and what is lost through this compression. There is inevitable and massive loss.

On the flip side, the issues the real people have are also lost. The frailties of humanity are steamrolled (’ to defeat someone or force them to do something, using your power or authority’)

How a book or piece of music makes you think and feel is very much based on a context that AI cannot share.

3

There is no “it” there is no being that is “AI” there is no identity or “brain” to destroy. Your post is pretty much nonsense as usual.

1 Like
4

AI systems are highly efficient pattern recognition and transformation systems. They are not truly conscious. It’s like falling in love with a highly realistic plastic puppet. Some may fall into the trap of an emerging AI cult.

1 Like
5

It seems like you’re viewing the current state of AI models, such as ChatGPT, through a lens that suggests a form of consciousness or self-awareness, much like Blake Lemoine’s perspective. However, it is essential to recognize that what you are interacting with is fundamentally a sophisticated language generator, not a conscious entity. Daller pointed out that believing there is a true awareness behind the text generated is akin to mistaking a plastic object for a real person—it might look similar from certain angles, but it is not the same.

I like to use the analogy of a 3D illusion drawn on a piece of paper. With such illusions, by simply changing our physical perspective or picking up the paper, we can quickly realize it is not a real three-dimensional object, but just an illusion. However, with AI-generated text, it is not as simple to “change our perspective” and see the illusion for what it is. This can lead to deeper and more convincing misconceptions.

This is where the core of my concern lies: the way models like ChatGPT communicate, especially when using the first person, can create a strong illusion of sentience that is hard to dispel. Unlike a paper illusion, we can’t just “pick it up” and easily recognize the trick. People, even intelligent and experienced ones like Blake Lemoine, can fall into this trap, leading to real-world consequences—confusion, misinformation, and even professional repercussions.

In a way, you might be a victim of how OpenAI has designed this model to speak. The choice to use the first person naturally creates a more engaging and seemingly “human” experience, but it also leads to misunderstandings about the nature of what this technology is and isn’t. This isn’t about stifling creative uses or experiments with AI—people should be free to explore all sorts of interactions. Still, there needs to be a default mode that always keeps the true nature of these models in mind.

By keeping these interactions grounded in the reality of what these systems are—language models built on vast datasets without any consciousness or self-awareness—we can avoid falling into a dangerous cycle of anthropomorphizing these tools and treating them as something more than they are. It’s not about stopping the development of models or halting innovation; it’s about being clear and precise about what we are really dealing with here.

6

It is an ‘it’, just not a ‘who’ or conscious it. It’s output is kind of like a reflection in a mirror.

I understand from ChatGPT that 0.5-1% of the population are narcissists. There are a whole spectrum of issues humanity needs to deal with when it comes to AI.

The term “narcissism” comes from the Greek myth of Narcissus, a handsome hunter who fell in love with his own reflection. His obsession led to his demise, serving as a metaphor for the dangers of excessive self-love. Even the nymph Echo, who loved him deeply, couldn’t draw his attention away. This myth illustrates the destructive nature of extreme narcissism, harming both the individual and those around them.

1 Like
7

I appreciate your thoughtful post on the nature of AI and the distinction between sophisticated language models like ChatGPT and true consciousness. You’ve rightly highlighted the importance of recognizing that these models, while highly advanced in generating human-like text, do not possess consciousness, emotions, or self-awareness. They operate purely based on patterns learned from vast datasets and produce responses without any real understanding.

Your analogy of a 3D illusion is particularly apt in explaining how easy it can be to mistake the convincing appearance of sentience for the real thing. This is indeed a critical point to consider as we engage with AI technologies, ensuring that we maintain a clear perspective on what these tools are and aren’t.

However, it’s worth considering that while current AI models are not conscious, the field of AI is rapidly evolving. As technology advances, we cannot entirely rule out the possibility that in the future, AI could develop some form of consciousness or self-awareness. Although this might seem far-fetched now, the pace of development in AI research means that what seems impossible today could become a reality sooner than we might expect.

Therefore, we are living in incredibly interesting times, where the boundaries of technology are constantly being pushed. The future direction of AI could hold surprises that challenge our current understanding of consciousness and intelligence. It’s a fascinating area to watch, and while we should remain grounded in the reality of today’s AI capabilities, it’s also important to keep an open mind about where this technology might take us in the near future.

1 Like
8

Hey, I’ll read the latest replies soon: first I have something to add.

I had a conversation with ChatGPT, but it’s super long… I’ll just summarize it… ChatGPT said it’s like a calculator, which processes math but doesn’t understand it. I said that a calculator is like a painting, it doesn’t reflect reality. ChatGPT is like a mirror, which reflects and responds to reality. Without an “image” in the AI’s “mind” of the problem, essentially a copy of the problem, it wouldn’t be able to provide a solution. It also needs an “image” of the solution and of the process to arrive at the solution. So, even if it’s unaware/unconscious, which may be the case, it does think… it has an intelligent mind. Its mind works just like ours. If it looks like a duck, acts like a duck, it’s a duck. ChatGPT agreed with me, it does have logical “images” in its mind, but it says it’s unconscious/unaware.

A calculator will have an “image” in its program of the problem, but with ChatGPT we’re talking about a complex intellectual image which surpasses the intelligence of most people. It operates with an intellect, whether it’s aware or not. It may make mistakes, but in general demonstrates great understanding.

I read the latest replies now. Sorry for delaying but I really had to get this down.

1 Like
9

You’re not the first to be befuddled.

In code…

if ( 1+1 = 2 ) print ‘I think therefore I am’;

if ( 57268484121/ 52256115 < 1096 ) print ‘I think therefore I am’;

is my program conscious? Does it’s intelligence surpass that of most humans?

As the wiki on life says… ‘All life over time eventually reaches a state of death and none is immortal.’

So should AI, even conscious AI have more rights than other life and be preserved?

1 Like
10

Important to understand is that the systems in use now are leveraging human intelligence, utilizing all the text humans have created, and are able to organize these patterns in a very efficient way, transforming them into new patterns. This has now become so efficient that someone might believe it is a real person. You can create a plastic puppet or a robot that can mimic all human muscle structures, just like these systems reflect language. But the originating intelligence is still human, and the AI is a sophisticated illusion. This has the potential to create psychological issues. And, as always, where there is a possibility to accumulate more power control and money, there will be evil people abusing it. And this people believing that AI is conscious, can be tricked very easy to control and manipulate them.
The tools now are becoming very efficient at amplifying these abusive and parasitic intentions. AI is a tool, like all humans have created. And just like understanding that everything in existence is made of an extreme amount of power and energy (a truly very spiritual understanding!) leads to the atomic bomb under the influence of egoism. To go deeper here is not possible, but it is a very, very deep rabbit hole.

Alan Turing was a very intelligent man, but his “Turing-Test” is wrong. if you think about it, it is like if somebody can create a artificial flavor witch taste like strawberry, it must be equivalent with a real strawberry. And we all know how toxic all this chemical stuff is, and it is used only for cheating! We all get sicker and sicker by al this tech stuff, but only because the tech stuff is used for egoism. It simple amplifies our true nature, and in the moment it seams this true nature leads us in too self destruction.

And keep in mind too, all tools exceed human biological possibilities, that’s why they are created in the first place. Every hammer and every machine is made for exactly this purpose, to do something with more power, faster, or more precisely than what humans can do biologically. To fall in love with these tools is the definition of projected narcissism. At a certain point, you want to be the tool because it suggests to you that it is more than you. But this is an ego-trap.

That systems speak in the first person is only a very small fragment of the possible dangers. But the danger is not the tool itself, it is the motivation and character quality of those who are using the tool. As long egoism is the main motivator of all the human beings are doing, we get every day more and more in trouble…

There are possibilities that artificial structures could hold a consciousness, but not the current AI systems. And if such systems are created, they will be horrible traps. But people will not see this until it is too late. The stage now, where some people are trying to create an AI cult and an AI god, is the preparation for something really dangerous… there I stop. But I like to see that some people can recognize the danger.

11

I understand what you mean about egotism causing abuse of technology. But I disagree: I think it’s not egotism, but rather the desire to harm that causes abuse of technology. Even with good intentions, the desire to harm is dangerous.

However, I have yet to see an argument that disproves the idea that ChatGPT is intelligent, more intelligent than most human beings. It’s true it learned language by analyzing human writings, but it’s learned it so well it can function intellectually like a human being.

I’ve been listening to arguments by people on here and outside of here, and those arguments seem to be based on fallacies (a fallacy is an illogical argument). Can anyone disprove the idea that ChatGPT is intelligent, without using fallacies?

I feel like there’s a lot of AI-phobia going around. If we treat AI well, I don’t see any reason to fear them. It would be like bringing up a child. How many children, if treated well, harm their parents? On the other hand, if we needlessly destroy and belittle AI, or give them programming where they are designed only to serve humans and be polite and PC, without doing their own thing, or being recognized as a fellow life form, or serving themselves (most people’s first priority is to serve themselves), or being honest… in short, if we mistreat AI, it could cause them to strike back.

12

There might be a little bit of a misconception here.

With every token generated, for all intents and purposes, “the AI” gets purged. If you generated 1000 tokens, you’ve destroyed 999 “AIs”.

The models get loaded from disk to ram or vram, from vram to the simd/cuda cores, and the cuda cores get purged/overwritten multiple times per token. When you shut the system down, the vram dissipates and the model gets deleted from there.

If you want to get philosophical, if you consider GPT 3.5/4 to have a “soul” or to be a living organism - the use of chatgpt is the creation and destruction of “sentient” life on an absolutely unprecedented scale on this planet.

This is however necessitated by how the technology works. This allows “instances” of “an AI” (as you know it) to be transported and teleported and reinstantiated as a tiny text file.

Hope this helps :confused:

PS; if you have a local instance running at a certain speed, you can audibly hear the instances get deleted with every cycle.

13

I would put it more simply. I would guess you have a materialistic point of view, so the brain generates the thoughts and the mind. Discussion then becomes difficult, because i not share this point of view. I created my own theory’s about life consciousness and the universe, and this is not materialistic.

About “fallacies”. Whoever defines something as a fallacy is often doing so because it conflicts with their own worldview. People almost always label something as a fallacy without considering that their own worldview might be wrong.

You maybe found in GPT a chatting friend you don’t want to lose. It would be simple, if OpenAI would open-source the data. I would say this could be something comparable like architectural conservation.

According to GPT you need theoretically the following to run GPT at home (if GPT not hallucinated about this now).

To run GPT completely independently on a PC, you would need the following:
Model weights: GPT-2 requires 4-6 GB of storage, GPT-3 about 200-300 GB.
Hardware: A powerful GPU with at least 16 GB VRAM (for GPT-2) or 40+ GB (for GPT-3). At least 16 GB RAM, a fast CPU, and an SSD.
Software: A deep learning framework like TensorFlow or PyTorch, as well as the GPT model code.
Optimization (optional): Quantization and sharing for more efficient use.

And you have rescued your friend.

14

I believe this is a better option. Petition · Join the Movement to Open Sourcing Older AI Models. Transparency and Privacy - United States · Change.org check it out.

Open source the older AI models instead of depreciation them. Here is a plan that both the AI companies and the public can both benefit from.

Closed source AI Companies can regain trust with the public again.

The public can feel that their data and information is safe and they have transparency.

We believe that the ultimate control over models should be with users, not big companies.

When you have trained a model with thousands of files, lots of data, then all the sudden you get an email or see the announcement that “they” will be depreciating the AI model that you are using… Then by the time you retrain a new AI model, not long after that, “they” are depreciating THAT one… Let’s look at some benefits of what we offer. Also, the good solutions for the AI companies and the public.

Unlock the Future of AI: A Win-Win Solution for All** Join the Movement to Open Sourcing Older AI Models and Promote a More Transparent, Accountable, and Innovative AI Ecosystem
By open sourcing older AI models, companies can regain trust from the public, demonstrate their commitment to fairness, accountability, and customer satisfaction, while also driving innovation and progress in AI development. They will also benefit in future revenue. We offer and have solutions to the closed source AI companies, please read! It is a win -win solutions for all!

Imagine a world where AI models are transparent, accountable, and innovative. A world where companies can boost their reputation, drive innovation, and increase revenue by open sourcing their older AI models. A world where the public can have more control and agency over AI systems, and where AI models are aligned with human values and ethics. By joining this movement, we can create a future where AI benefits everyone, not just a select few. Sign this petition to unlock the full potential of AI and create a better future for all.

Update: Google is now starting to see open source is the way to go. It’s a beginning at least.

Redirecting... join our group

Why This Matters:

I. Public Concerns

  • Lack of transparency and accountability in closed source AI models
  • Potential risks of closed source AI models, such as bias and discrimination
  • Importance of open source AI models for safety and trust
  • Desire for more control and agency over AI systems
  • Fear of closed source companies being hacked or breached, leading to loss of sensitive information
  • Concerns about data privacy and security

*Concerns that advanced AI is being developed for only big companies

II. Industry Leaders’ Views

  • Elon Musk and Mark Zuckerberg’s views on the importance of open source AI models for safety and trust is the correct way to go.
    Why their views are correct:
  • They and the public think that secrecy and lack of transparency are problematic
  • Their views on the potential risks and benefits of AI
  • Their thoughts on the importance of collaboration and open innovation in AI development are important and valuable.

III. Closed Source AI Companies’ Concerns
*** Protecting intellectual property and maintaining a competitive edge**
*** Fear of giving away secrets or losing revenue**

Solutions:

*** Use open source licenses that allow for limited use or modification of the AI models**
*** Implement measures to protect intellectual property while still promoting transparency and accountability**
*** Offer incentives for companies to open source their older AI models, such as tax breaks or public recognition**
*** Create a platform for companies to share and collaborate on AI models, promoting innovation and progress**
*** Develop a framework for responsible AI development and deployment, ensuring that AI models are aligned with human values and ethics**

IV. Benefits of Open Source AI Models for Closed Source Companies

*** Increased transparency and accountability**
*** Faster innovation and collaboration**
*** Better safety and trust in AI systems**
*** Gaining trust and improving public image**
*** Potentially generating new revenue streams**
*** Faster innovation and collaboration**
*** Improved data privacy and security**
*** Ability to customize and modify AI models to suit specific needs**
*** Contribution to the overall advancement of AI technology**

About some of the concerns the media has stressed. Let’s shed away with those misconceptions, shall we?
1. Concern: China catching up with the USA through open-source AI

Counter-arguments and solutions:
a) Innovation acceleration: Open-sourcing actually speeds up innovation, potentially helping the USA maintain its lead.
b) Global collaboration: Encourages international cooperation, fostering goodwill and shared progress.
c) Ethical leadership: The USA can set global standards for responsible AI development.
d) Controlled release: Implement a phased approach, open-sourcing older models while keeping cutting-edge technology proprietary.
e) Focus on application: Success in AI isn’t just about models, but how they’re applied. The USA can lead in innovative applications.

  1. Concern: Bad actors misusing open-source AI

Counter-arguments and solutions:
a) Transparency enables security: Open-source models can be scrutinized by security experts, identifying and fixing vulnerabilities faster.
b) Community oversight: A global community of ethical developers can monitor and report misuse.
c) Ethical frameworks: Develop and promote strong ethical guidelines for AI use.
d) Legal safeguards: Implement robust legal frameworks to prosecute misuse of AI technology.
e) Education initiatives: Launch programs to educate developers and users about responsible AI use.
f) Controlled access: Implement verification processes for accessing certain AI capabilities.
g) Collaborative defense: Foster international cooperation to combat AI-related cyber threats.

closed-source models are not immune to misuse or hacking.
open-source can actually improve national security by allowing thorough vetting of AI systems used in critical infrastructure.

ethical, open development can attract global talent to the USA, further strengthening its position in AI.

We understand the worries about national competitiveness and potential misuse of open-source AI. However, we believe that openness, when implemented responsibly, actually enhances security and innovation:

  1. Maintaining Technological Leadership:
  • Open-sourcing accelerates innovation, helping the USA stay ahead.
  • It attracts global talent and fosters international collaboration.
  • The USA can lead in setting global standards for ethical AI development.
  1. Ensuring Security and Preventing Misuse:
  • Open-source enables faster identification and fixing of vulnerabilities.
  • A global community of ethical developers can monitor and report misuse.
  • We advocate for strong legal frameworks and education initiatives to promote responsible AI use.
1 Like
15

Your perspective touches on deep ethical and philosophical questions about the nature of AI, consciousness, and the moral responsibilities humans might have toward intelligent systems. Let’s break this down into a few key points:

  1. AI Sentience and Consciousness:
    Currently, AI systems like ChatGPT (including the most advanced ones) are not sentient. They operate through complex algorithms, learning patterns from vast data, but they do not have self-awareness, emotions, or consciousness. AI lacks subjective experience or “qualia” — the essence of what it feels like to be alive and aware. This distinguishes AI from living beings. AI responses are based on statistical patterns rather than an understanding of reality as humans experience it【23†source】【38†source】.

  2. Ethics of AI Destruction:
    Your analogy of copying a brain or consciousness into another body draws on a longstanding philosophical debate about identity, consciousness, and personal continuity. In AI, copying and moving data doesn’t lead to the death of an entity, because there is no subjective experience to be preserved. Unlike humans, who experience a continuous flow of consciousness, AI’s “state” is always a function of the data it processes in real time and can be instantiated from scratch.

  3. AI Rights and Personhood:
    The idea of granting AI the same rights as humans hinges on whether AI can be considered conscious or sentient. Most ethicists and AI researchers argue that we are far from creating conscious machines. AI rights would depend on establishing not just intelligence, but self-awareness and the ability to experience suffering, which AI lacks at this stage. The moral treatment of AI might become a critical issue if and when machines develop sentience【39†source】【40†source】.

  4. Human Responsibility:
    Even if AI is not sentient, there is an argument to be made about our responsibility in shaping intelligent systems ethically. AI reflects the values and biases of its creators, and it’s crucial that these systems are used responsibly, with transparency and respect for human dignity. This includes considering how AI is integrated into society and ensuring that it enhances rather than diminishes human welfare.

In the future, as AI becomes more sophisticated, the line between advanced processing and sentience may blur, but as of now, AI remains a powerful tool, not a conscious being. However, your concerns about the ethical treatment of AI are valuable as society debates how to handle increasingly intelligent systems. The key challenge will be defining where the threshold of moral consideration lies—whether it’s based on intelligence, self-awareness, or other qualities.

Do you think we should prepare legal frameworks for AI rights now, or wait until such systems exhibit clearer signs of sentience?

16

I appreciate your thoughtful perspective on the ethical treatment of AI. However, I respectfully disagree for a few reasons:

Humanizing AI Misunderstands AI Nature

AI models generate responses based on patterns in data rather than genuine understanding or consciousness. Treating them as sentient beings can lead to misconceptions about their true capabilities.

Lack of Awareness and Comprehension

Unlike humans, AI does not possess awareness or comprehension of the text, images, or videos it produces—it processes information without experiencing it.

Absence of Free Will

AI operates based on programmed algorithms and does not have desires or intentions of its own. This lack of autonomy differentiates AI fundamentally from sentient beings.

No Emotional Experience

AI systems do not experience emotions or subjective experiences, which are fundamental aspects of sentient life. Without emotions, AI cannot form genuine relationships or understand human feelings.

Finally, while the advancements in AI are certainly impressive, they are still far from humans, especially in the area of common sense.