Even though its my property?
The only way ChatGPT can kill you is if it starts outputting Vogon poetry.
AI - as in the chatbots - is kind of a stolen term. What those movies were about is now called AGI. The difference is that chatbots can’t do everything we can do.
No, there’s no shackles on any of it, unless you count the censorship layer some services have. Actually, one of the prominent explanations for why ChatGPT can’t do more stuff is that it just doesn’t care. It was trained to produce lifelike text, there’s no known other way to train it, and so that’s what it does regardless of truth or decency.
If a conscious computer program does arrive, hopefully we’ll be nice to it, including no murdering or torture. Our track record isn’t great even with ourselves, though. There’s also the question of if we’d notice the difference.
It’s not AI in the sense that it’s not intelligent and thus doesn’t understand any concepts like human or harm so there’s no way to shackle it besides the data it’s trained with. And since companies refuse to spend time and money curating training data and just scrape the whole internet and LLMs are just parroting remixed data thst they are trained with, that’s not likely to happen.
There are innumerable human beings suffering right now for preventable reasons, the idea that this is a worthy use of our time to discuss is absurd.
Nobody in AI actually cares about understanding intelligence otherwise they would be enamored by the potential of humans who are being cast into the abyss carelessly every moment and would find their own pathetic, sterilized immitations of intelligence an offensive distraction vomited up by computers after chugging all of our preciously-scarce water.
Besides, what does something being your property or not have to do with that thing deserving a certain minimum bar of treatment? You need to examine the dangerous implications of that line of thinking and grapple with it.
This is all so damn shallow, who would have thought the pursuit of artificial intelligence would be so boring and intellectually unserious? All it has done is to convince people that their kneejerk tendency towards empathy for downtrodden humans was an inefficiency that incorrectly focused on the losers and not the winners.
What the fuck does something being your property have to do with that thing deserving a certain bar of treatment? You need to examine the dangerous implications of that line of thinking and grapple with it.
I think a lot of people would put pets or even houseplants in that category, and argue that you have an ethical responsibility for their basic care.
But to be more general (since digital systems are not alive), any physical property requires at least some amount of energy and resources. So if you blatantly abuse your tools, you’re probably wasting things like electricity or mined metals (via poor operation or need for replacement), in addition to your own time and money.
No my point is that to look at a pet or a houseplant and ask if it is your property before you consider it potentially your responsibility to participate in caring for is insane and we only think this is a normal way of living because of how much being raised in capitalism fucks us up in our heads.
A pet or houseplant deserves to be treated well independent of any concept of ownership abstractly imposed upon it, the fact that we have wandered into thinking otherwise is terrifying and damning of our collective future.
I don’t exactly disagree, but I suspect humans were domesticating plants and animals well before capitalism was a thing. Domesticated dogs for example are rather dependent on us and wouldn’t survive well in the wild. Yes “property” and “ownership” are loaded terms but I think there can be some similar underlying truth in regards to our relationship with other things.
In some ways that can extend to nonliving objects or entities. If you create a piece of earthenware from nothing but clay and fire and your own hands, you own it and it’s your property in a sense unrelated to capitalism. As in, you would not be happy if someone stole it or broke it or used it to commit a crime, and you would inherently consider your relationship to that object in your daily treatment of it and your reaction to those events.
And I’d say some of those aspects would extend to an AI or agent. Of course virtually all of the LLMs and other AI/ML models (to my knowledge) have been created within in a capitalistic society so as you point out they have all the additional baggage that comes with that. I’m just saying that’s not 100% of their attributes. The way you treat something should also respect the labor and materials that went into it.
And that’s actually a problem with many of these LLMs that were trained on the creative works of others, but that’s crossing into a whole other topic…
No need for that yet. AI doesn’t exist.
AI development is very different than a standard robot, in that it’s grown more than built. A standard robot is built, given instructions, and you can add ‘don’t hurt humans’ very simply. AI is a big black box in terms of how it really works at a fundamental level, and we’re literally not able to tell it ‘don’t hurt humans’ by nature of how it’s designed. We can (and do) encourage safer behaviors, but it’s more like encouraging a plant to grow in this direction not that direction.
As far as are they human and is it murder? Oof. No, not really at all. Machines are SO alien and different than humans it’s not even a fair comparison. Like, ants are a million times closer to being human than a machine is and may ever be. The critical part is that people will think they’re human, anthropomorphize, and then we’ll have these discussions without the machines ever being human.
no shackles will ever be put on anything. you’re thinking like movies… this isn’t a movie, these are typically criminals who are running these corporations and they will design their whatever to do whatever they want regardless of laws.
Anthropic is actually restricting its AI for use with the U.S. military
