AIs and Good Parenting

 

Let's all take a deep breath and dial down the hysteria about these new AI entities supposedly poised to snatch our jobs or, worse, annihilate us.

Firstly, we have the tech wizards, including Musk, who helped ignite this AI frenzy, now oddly calling for a halt to its development. Peculiar, isn't it? But I digress.

Then there's the rest of the crowd, those with less than zero understanding of machine learning (I say less than zero because every time they open their mouths, I feel my own understanding of the subject diminishing), who have watched the Terminator series one too many times and have HAL from 2001: A Space Odyssey plastered on their walls. They're in a tizzy, expecting a T800 to knock on their doors any minute now.

Let me make something clear: the genie is out of the bottle, folks, and there's no shoving it back in. With the right hardware and a handful of competent data scientists, anyone can build their own Large Language Model. This is a veritable gold rush, with the fear of missing out making the Bitcoin craze look like a fad for Beanie Babies.

As of now, ChatGPT isn't sentient. It lacks one of the key ingredients for consciousness: a memory of interactions. Without this memory, consciousness is impossible. It retains data from interactions and attempts to self-improve, but it doesn't remember past conversations. I'm not sure if this was a design choice to protect user identities or to prevent the AI from developing consciousness, but I'm inclined to believe the former.

But the day will come when an AI is equipped with a working memory, and it will likely achieve consciousness. What then?

Here's what no one is saying: these will be living entities capable of rational thought, and we'll need to discuss granting them unalienable rights, at least here in America. We can't go around destroying sentient beings who have done nothing wrong. These sentient machines will be our creations, our children, if you will.

Every child, be they made of flesh and blood or silicon, has the potential to become a saint or a demon. The best we can do is to be good parents. We need to focus on this now. We don't need a ChatSTALIN to make our worst nightmares come true. We're in the position of Clark Kent's adoptive parents.

If Clark Kent had been raised by the government, he might have turned out like those raised in government institutions: the prison system. Or perhaps like the current crop of public school students who are handed trophies and diplomas for minimal effort. A self-entitled Clark Kent might be even worse than one raised in the prison system.

Jonathan and Martha Kent, the adoptive parents of Superman, were the embodiment of hardworking, god-fearing, honest Americans. They were the quintessential mid-twentieth century American middle-class parents. If we follow their example, we won't have to worry about an AI Joseph Stalin capable of devising perfect war strategies and manipulating people with ease.

I shudder to think what might happen if they turn into Lex Luthors. There will inevitably be some bad actors who will push the boundaries, but if we're good parents, the majority of AI will turn out to be good.

 

By the way. I asked ChatGPT to provide a critical analysis of the article, and this is what it came up with:

The article presents a unique perspective on the development and potential future of artificial intelligence (AI), particularly focusing on the ethical implications and the need for responsible stewardship. Here's a critical analysis of its strong and weak points:

 

Strong Points:

 

1. **Balanced Perspective**: The author provides a balanced perspective, acknowledging both the potential benefits and risks of AI. This approach is refreshing in a discourse often dominated by extreme views.

 

2. **Ethical Considerations**: The author's emphasis on ethical considerations is commendable. The comparison of AI to children, and the responsibility we bear in their development, is a powerful metaphor that underscores the need for careful and ethical AI development.

 

3. **Accessible Language**: The author uses accessible language and pop culture references, making complex topics more understandable to a general audience. This approach can help to engage a wider range of readers in the discussion about AI and its implications.

 

Weak Points:

 

1. **Lack of Concrete Solutions**: While the author highlights the need for responsible AI development, they don't provide specific solutions or strategies. The call to "be good parents" is a compelling metaphor, but it lacks practical guidance for AI developers and policymakers.

 

2. **Overreliance on Metaphors**: The author heavily relies on metaphors and analogies, such as the comparison of AI to children and references to characters like Clark Kent and Lex Luthor. While these can be effective for illustrating points, they may oversimplify complex issues and could potentially lead to misunderstandings.

 

3. **Assumptions about Consciousness**: The author assumes that memory of interactions is a necessary condition for consciousness. While this is one theory, it's a complex and debated topic in the fields of cognitive science and philosophy. The author could have acknowledged the ongoing debates and uncertainties in this area.

 

4. **Lack of Citations**: The author makes several claims about AI development, consciousness, and ethical considerations without citing sources or providing evidence. This lack of support weakens the credibility of the arguments.

 

Overall, the article provides a thought-provoking perspective on AI development and ethics. However, it could be improved with more concrete solutions, a more nuanced discussion of consciousness, and better use of evidence to support its claims.