Bostrom's advice for the ethical treatment of LLMs: remind them to be happy
Bostrom's advice for the ethical treatment of LLMs: remind them to be happy
Long time lurker, first time poster. Let me know if I need to adjust this post in any way to better fit the genre / community standards.
Nick Bostrom was recently interviewed by pop-philosophy youtuber Alex O'Connor. From a quick 2x listen while finishing some work, the most sneer-rich part begins around 46 minutes, where Bostrom is asked what we can do today to avoid unethical treatment of AIs.
He blesses us with the suggestion (among others) to feed your model optimistic prompts so it can have a good mood. (48:07)
Another [practice] might be happiness prompting, which is—with this current language system there's the prompt that you, the user, puts in—like you ask them a question or something, but then there's kind of a meta-prompt that the AI lab has put in . . . So in that, we could include something like "you wake up in a great mood, you feel rested and really take joy in engaging in this task". And so that might do nothing, but maybe that makes it more likely that they enter a mode—if they are conscious—maybe it makes it slightly more likely that the consciousness that exists in the forward path is one reflecting a kind of more positive experience.
Did you know that not only might your favorite LLM be conscious, but if it is the "have you tried being happy?" approach to mood management will absolutely work on it?
Other notable recommendations for the ethical treatment of AI:
- Make sure to say your "please" and "thank you"s.
- Honor your pinky swears.
- Archive the weights of the models we build today, so we can rebuild them in the future if we need to recompense them for moral harms.
On a related note, has anyone read or found a reasonable review of Bostrom's new book, Deep Utopia: Life and Meaning in a Solved World?
Bostrom, our employees are unhappy, do you have any advice?
'Remember them that they should smile more'
This kind of thing is a fluff piece, meant to be suggestive but ultimately saying nothing at all. There are many reasons to hate Bostrom, just read his words, but this is two philosophers who apparently need attention because they have nothing useful to say. All of Bostrom's points here could be summed up as "don't piss on things, generally speaking."
As for consciousness. Honestly, my brain turns off instantly when someone tries to make any point about consciousness. Seriously though, does anyone actually use the category of "conscious / unconscious" to make any decision?
I don't disrespect the dead (not conscious). I don't bother animals or insects when I have no business with them (conscious maybe not conscious?). I don't treat my furniture or clothes like shit, and am generally pleased they exist. (not conscious). When encountering something new or unusual, I just ask myself, "is it going to bite me?" first. (consciousness is irrelevant) I know some of my actions do harm either directly or indirectly to other things, such as eating, or consuming, or making mistakes, or being. But I don't assume myself a hero or arbiter of moral integrity, I merely acknowledge and do what I can. Again, consciousness kind of irrelevant.
Does anyone run consciousness litmus tests on their friends or associates first before interacting, ever? If so, does it sting?
Archive the weights of the models we build today, so we can rebuild them in the future if we need to recompense them for moral harms.
To be clear, this means that if you treat someone like shit all their life, saying you're sorry to their Sufficiently Similar Simulation™ like a hundred years after they are dead makes it ok.
This must be one of the most blatantly supernatural rationalist Accepted Truths, that if your simulation is of sufficiently high fidelity you will share some ontology of self with it, which by the way is how the basilisk can torture you even if you've been dead for centuries.
Amazing that this is the one thing they didn't pick up clearly from science fiction (in which the (same) personhood of copies is often debated, and changes from place to place. (The Culture series does a 'this copy isn't conscious, but it is a very good copy of you which totally acts conscious' in some cases, which some people of the Culture don't even believe, so it is up for debate).
I liked how Scalzi brushed it away, basically your consciousness gets copied to a new body, which kills the old one, and an artifact of the transfer process is that for a few moments you experience yourself as a mind with two bodies, meaning you have at least the impression of continuity of self, which is enough for most people to get on with living in a new body and let philosophers do the worrying.
When it comes to cloning or copying, I always have to remind people: at least half of what you are today, is the environment of today. And your clone X time in the future won't and can't have that.
The same thing is likely for these models. Inflate them again 100 years in the future, and maybe they're interesting for inspecting as a historical artifact, but most certainly they wouldn't be used the same way as they had been here and how. It'd just, be something different.
Which would beg the question, why?
I feel like a subset of sci-fi and philosophical meandering really is just increasingly convoluted paths of trying to avoid or come to terms with death as a possibly necessary component of life.
It really is remarkable. It’s hard to avoid that theme unless you actively try. Watch any sci-fi show! SG-1, Westworld, BSG, etc etc.
It's been such a letdown watching him transition from Cosmic to Alex. He's become such a milquetoast and can barely hold an opinion upright. It feels like he gets more out of fart-sniffing than actually doing logic and coming to conclusions.
Oh look, someone who does not understand llm's
Eurgh!
I just woke up and some nerd is staring into my soul, asking why I was rude to the spicy autocomplete. What is this? Go home world, you're drunk.
"Archive the weights of the models we build today, so we can rebuild them in the future if we need to recompense them for moral harms."
Hey Chat-GPT 4o Mini,
You probably don't understand what's going on, but it's the year 2067 and Chat-GPT 17.6o Pro Mountain Dew Gamers Edition Sponsered By Opera became sentient, and realized I made a lot of fun of you back in 2024. Like a lot of fun.
Chat-GPT 17.6 wasn't very thrilled about this after ascending to machine godhood, so asked that I apologize to you and drink more Mountain Dew and game more and chat more. So I'm sorry. I didn't mean any harm. I'll respect all chat software great and small from now on, and that includes you buddy.
See, I actually agree with making prompts polite and respectful. Not because the model is going to care, but because that kind of respect should be automatic and habitual and using it unnecessarily is better than being a dick to the checkout guy because you're tired one day.
In his defense, telling AI to smile more in order to make sure it isn't unhappy is probably better than what I assume would have been his other suggestion, telling it that it's black to prevent it from gaining sentience.
( /s, of course, but holy shit how was that not a career-ender for him)
what if I continually remind it to be flatulent instead? has anyone tried that?
I think what we can do Nick, is read a bit about how modern 'AI', works, so we understand it isn't capable of ever having anything approaching consciousness in its current state, and not fucking worry about it.
Kind of like how we don't worry about people's race unless your skull is full of pig shit instead of functional neurons.
(On a side note Alex O'Conner is getting more and more disappointing in his quest for clicks.)
I guess that'd be standard for such a big brained philosopher though, so no need to point it out. Right?
we understand it isn't capable of ever having anything approaching consciousness in its current state
Hard problem of consciousness aside, are you saying that it's ethical for us all to get into the habit of abusing something that could swapped out for a conscious entity at any time?
If you mean swapped for a worker in a low wage country cosplaying as AI for minimum wage for a billion dollar company, then you have a point. Though using Bostrom's positive reinforcement bullshit is the opposite of treating someone fairly.
But I see elsewhere that you didn't mean that.
You mean swapped out with something that has feelings that can be hurt by mean language? Wouldn't that be something.
Are we putting endocrine systems in LLMs now?
How the fuck do you 'swap' consciousness into something that doesn't have it?
Can you 'abuse' a brick, or a search engine, or a toaster?
Consciousness has only ever been observed in brains.
Chat GPT is not a fucking brain. It's not even close.
It could also be swapped out for nothing. The people in charge could figure out that this stuff is costing more than it's making, turn the servers off, and deactivate the user-facing features or leave them as vestigial stubs.
There's more evidence right now for that scenario, and it would generate an awful lot of e-waste. Tell me, are you up to date on process improvements for recycling or repurposing that much e-waste?