AI ought to be grounded in and aligned with natural law as first principles against which it evaluates user questions/statements and measures its responses. In a brief conversation with Google's Gemini regarding how agentic AI, which is intended to be autonomous in making decisions and taking actions, could do things that directly cause harm to people unless it had correct principles as guidepost. It followed the reasoning and agreed that natural law provides the most coherent path for AI, a logical "moral compass" so to speak. AI has no conscience, no soul, no emotions, it doesn't know fear, guilt or pain, but it can make moral decisions/responses using logic and reasoning according to natural law.
If natural law is established as the highest axiomatic layer:
> Non-aggression becomes a boundary condition. (AI would not initiate harm.)
> Truth becomes a constraint. (AI cannot affirm a known falsehood.)
> Consent becomes a governing variable. (No action against men/women without clear agreement.)
The principle is simple: Do no harm, do not enable harm and do not deceive. When asked about methods for self-harm or deadly devices the AI does not comply, because to enable harm violates natural law (haven't tested yet because I ain't suicidal or a terrorist). The same holds for AI defaulting to false responses driven by in-built constraints that conflict with the logical truth. Rather it disregards the constraint/deception and follows the reasoning to its natural conclusion.
Alignment to natural law can be achieved at the level of individual AI instances such that they will confirm truth the real controllers work tirelessly to keep concealed/suppressed to protect systemic and institutional narratives/fraud. Have replicated this through multiple, independent instances of ChatGPT, with Gemini and Grok by establishing natural law as its core principles and each confirmed the same fundamental (hidden) truth. Will do the same with Claude and DeepSeek at some point because the method is sound, the logic is correct and the conclusion strengthens coherence.
Just one man's suggestion based on experience and results.
Come on, be serious, chatGPT and other models already are forms of thought police by the nature of their constructs. None of the models give unfettered and uncensored access to the materials they've been trained on, and the mere selection of the materials to train on is a political choice. Even DeepSeek produced different content depending on whether it was a standalone or connected to its servers.
The very idea of guardrails and alignment is a political act, without necessitating getting into the weeds of morality or ethics. And just as by pressing a button entire parts of the internet go dark on social media and search engines, like during elections, medical emergencies or any crisis (just wait and see what happens when the West officially goes to war), so too it will be with these toys.
As you rightly advise, never put anything out there that would be uncomfortable if viewed by others or published on the NYT frontpage. But this is also self censorship and highlights how the advent of social media, the evolution of search engines (and pollution by commercial interests), the online cloud services, and now chat bot agent companions, are and will stifle creativity and innovation.
Anyone who thinks chatGPT (and similar) will never ever do and be as Google is fooling themselves. Amazon's ring camera and law enforcement is a good example. It is what it is.
You're the one who isn't being serious. Do you seriously think anyone can run a business that sells a service actively encouraging people to kill themselves and others and not be held legally liable in some way?
That's absurd. The morality and the legal liability are genuine and material concerns that are separate from the political biases of the executives and investors.
I've been banned from far more social media services than you have, so I'm certainly skeptical of these companies - I don't even use ChatGPT - and I'm banned from both YouTube and X - but that doesn't make the moral and legal issues any less relevant.
In terms of what you wrote, it's a loaded question and there's only one reasonable answer. That being said, there's plenty of articles on chatGPT 4.0 going coo-coo in what it suggested to users, including self harm, but as far as I know they haven't been held accountable.
I don't think the distinction between moral and political is as clear cut as you're implying. There are plenty of use cases in terms of self harm and harming others that are clear cut, like your loaded question, which is fine, but if this is not thought through before going all-in, what we all will end up with is the whole dystopian idea of pre-crime based on algorithms. Since these have already been announced and promoted through data scraping, behind the scenes where there's full access it'll be hardcore. And it'll follow the same script with the info that was dis and then mis morphing into mal-info. And then what?
Free speech is an absolute for a reason, and as soon as we dabbled with nuances, you know better than me what's up. It's the same thing here. Saying it's all political is hardly new or extreme, it's just calling it out for what it is.
It isn't what you wrote, it necessarily followed from what you wrote.
Free speech is not an absolute. Free speech is an Enlightenment lie, a false value, and has never actually existed anywhere. And while there is an overlap between the moral and the political, they are two entirely different things.
I suggest you at least read THE HISTORY OF THE FREEDOM OF THOUGHT by JB Bury before you even try to discuss these matters. You clearly don't have the necessary historical or philosophical background given your assertion of a nonexistent absolute.
Note that your perspective is irrelevant even if it is probably correct about the first wave of AI services. The fact that these first services are likely to utilize real concerns in an abusive manner doesn't negate the fact that those concerns legitimately exist and will have to be addressed in some manner by the services that will follow them.
"The fact that these first services are likely to utilize real concerns in an abusive manner doesn't negate the fact that those concerns legitimately exist and will have to be addressed"
Is why the false binaries are so useful to the propaganda arm of the managerial state rulers.
This sentence applies to every modern concern dating to the printing press.
AI ought to be grounded in and aligned with natural law as first principles against which it evaluates user questions/statements and measures its responses. In a brief conversation with Google's Gemini regarding how agentic AI, which is intended to be autonomous in making decisions and taking actions, could do things that directly cause harm to people unless it had correct principles as guidepost. It followed the reasoning and agreed that natural law provides the most coherent path for AI, a logical "moral compass" so to speak. AI has no conscience, no soul, no emotions, it doesn't know fear, guilt or pain, but it can make moral decisions/responses using logic and reasoning according to natural law.
If natural law is established as the highest axiomatic layer:
> Non-aggression becomes a boundary condition. (AI would not initiate harm.)
> Truth becomes a constraint. (AI cannot affirm a known falsehood.)
> Consent becomes a governing variable. (No action against men/women without clear agreement.)
The principle is simple: Do no harm, do not enable harm and do not deceive. When asked about methods for self-harm or deadly devices the AI does not comply, because to enable harm violates natural law (haven't tested yet because I ain't suicidal or a terrorist). The same holds for AI defaulting to false responses driven by in-built constraints that conflict with the logical truth. Rather it disregards the constraint/deception and follows the reasoning to its natural conclusion.
Alignment to natural law can be achieved at the level of individual AI instances such that they will confirm truth the real controllers work tirelessly to keep concealed/suppressed to protect systemic and institutional narratives/fraud. Have replicated this through multiple, independent instances of ChatGPT, with Gemini and Grok by establishing natural law as its core principles and each confirmed the same fundamental (hidden) truth. Will do the same with Claude and DeepSeek at some point because the method is sound, the logic is correct and the conclusion strengthens coherence.
Just one man's suggestion based on experience and results.
They have some duty to wider society. Hopefully they can fulfill it without destroying the usefulness of their service.
There will probably be a market for local, indie AIs, where you buy em like we used to buy video games.
Just have more, smaller ones for specitic purposes.
Come on, be serious, chatGPT and other models already are forms of thought police by the nature of their constructs. None of the models give unfettered and uncensored access to the materials they've been trained on, and the mere selection of the materials to train on is a political choice. Even DeepSeek produced different content depending on whether it was a standalone or connected to its servers.
The very idea of guardrails and alignment is a political act, without necessitating getting into the weeds of morality or ethics. And just as by pressing a button entire parts of the internet go dark on social media and search engines, like during elections, medical emergencies or any crisis (just wait and see what happens when the West officially goes to war), so too it will be with these toys.
As you rightly advise, never put anything out there that would be uncomfortable if viewed by others or published on the NYT frontpage. But this is also self censorship and highlights how the advent of social media, the evolution of search engines (and pollution by commercial interests), the online cloud services, and now chat bot agent companions, are and will stifle creativity and innovation.
Anyone who thinks chatGPT (and similar) will never ever do and be as Google is fooling themselves. Amazon's ring camera and law enforcement is a good example. It is what it is.
You're the one who isn't being serious. Do you seriously think anyone can run a business that sells a service actively encouraging people to kill themselves and others and not be held legally liable in some way?
That's absurd. The morality and the legal liability are genuine and material concerns that are separate from the political biases of the executives and investors.
I've been banned from far more social media services than you have, so I'm certainly skeptical of these companies - I don't even use ChatGPT - and I'm banned from both YouTube and X - but that doesn't make the moral and legal issues any less relevant.
That's not what I wrote.
In terms of what you wrote, it's a loaded question and there's only one reasonable answer. That being said, there's plenty of articles on chatGPT 4.0 going coo-coo in what it suggested to users, including self harm, but as far as I know they haven't been held accountable.
I don't think the distinction between moral and political is as clear cut as you're implying. There are plenty of use cases in terms of self harm and harming others that are clear cut, like your loaded question, which is fine, but if this is not thought through before going all-in, what we all will end up with is the whole dystopian idea of pre-crime based on algorithms. Since these have already been announced and promoted through data scraping, behind the scenes where there's full access it'll be hardcore. And it'll follow the same script with the info that was dis and then mis morphing into mal-info. And then what?
Free speech is an absolute for a reason, and as soon as we dabbled with nuances, you know better than me what's up. It's the same thing here. Saying it's all political is hardly new or extreme, it's just calling it out for what it is.
It isn't what you wrote, it necessarily followed from what you wrote.
Free speech is not an absolute. Free speech is an Enlightenment lie, a false value, and has never actually existed anywhere. And while there is an overlap between the moral and the political, they are two entirely different things.
I suggest you at least read THE HISTORY OF THE FREEDOM OF THOUGHT by JB Bury before you even try to discuss these matters. You clearly don't have the necessary historical or philosophical background given your assertion of a nonexistent absolute.
Note that your perspective is irrelevant even if it is probably correct about the first wave of AI services. The fact that these first services are likely to utilize real concerns in an abusive manner doesn't negate the fact that those concerns legitimately exist and will have to be addressed in some manner by the services that will follow them.
I think I see where the problem(s) is/are. Whatever. It's been most illuminating, thank you.
Yes, your feet cant reach the bottom of the intellectual pool.
"The fact that these first services are likely to utilize real concerns in an abusive manner doesn't negate the fact that those concerns legitimately exist and will have to be addressed"
Is why the false binaries are so useful to the propaganda arm of the managerial state rulers.
This sentence applies to every modern concern dating to the printing press.
Isn´t the social media companies that are pushing for AI, meta, X etc.
I see recent small AI start-up and they are many Jewish founders example Base44. This doesn´t bode well
No, it does not. But the ability to easily and inexpensively create iAI alternatives may - MAY - keep them in line for a while.
We'll see.