I don’t actually know what to do here. But I do honestly think this is likely to be representative of a pretty big problem. I don’t know that any particular regulation is a good solution.
Just to go over the obvious points:
I definitely expect AI partners to at least somewhat reduce overall fertility of biological humans, and to create “good enough but kinda hollow” relationships that people err more towards on the margin because it’s locally convenient even among people who I wouldn’t normally consider “suffering-ly single”
Yes, there are real suffering people who are lonely and frustrated, who’s lives would be definitively better with AI romantic partners
Regulations are a blunt, dumb instrument, and I think there are a lot of good reasons to be skeptical of paternalistic laws (i.e. Prohibition and the War on Drugs seem like a failures that made things worse)
I think it’ll be really hard to regulate this in any kind of meaningful way. This is much harder to regulate than drugs or alcohol.
My guess is that porn and romance novels/movies have a mixture of effects that include “making sexually frustrated young men less likely to rape or be pushy about sex”, but also setting unrealistic expectations that subtly warp relationships.
The era of “human fertility” is really just a blip here – within another couple decades (at most 50 years) after most after AI partners become threatening, I’d bet moderately heavily on us having uploads, and full fledged artificial life being pretty common, and pretty quickly dominating the calculus of what’s at stake. (there might or might not still be biological humans, I’d probably bet against biological humans lasting more than 200 years, but they’ll be a small fraction of sentient life that morally matters)
With all that in mind, what is the right thing to here? Hell, I do not know. But rounding this off whatever your preferred simple ideological handle is, is going to massively fail at protecting the future. Figuring out the answer here involves actual math and weighing tradeoffs against each other for real.
I think it’ll be really hard to regulate this in any kind of meaningful way. This is much harder to regulate than drugs or alcohol.
Why do you think so? Even if open-source LLMs and other necessary AI models will develop so quickly that real-time interaction and image generation will be possible on desktops, and there will be open-source waifu projects, mobile is still controlled by app stores (Apple and Google Play) where age-based restrictions could be easily imposed. Jailbreaking/unlocking the phone (or installing an apk from Github), connecting it to a local server with AI models, and maintaining the server is so burdensome that I expect 90+% of potential users of easy-to-access AI partner apps would fall off.
This is not to mention that open-source projects developed by hardcore enthusiasts will probably cater for their specific edgy preferences and not appeal to a wide audience. E.g., an open-source project of anime waifus may optimise to trigger particular fetishes rather overall believability of the AI partner and the long-term appeal of the “relationship”. Open-source developers won’t be motivated to optimise the latter, unlike AI partner startups, because their lifetime customer value would directly depend on that.
I don’t actually know what to do here. But I do honestly think this is likely to be representative of a pretty big problem. I don’t know that any particular regulation is a good solution.
Just to go over the obvious points:
I definitely expect AI partners to at least somewhat reduce overall fertility of biological humans, and to create “good enough but kinda hollow” relationships that people err more towards on the margin because it’s locally convenient even among people who I wouldn’t normally consider “suffering-ly single”
Yes, there are real suffering people who are lonely and frustrated, who’s lives would be definitively better with AI romantic partners
Regulations are a blunt, dumb instrument, and I think there are a lot of good reasons to be skeptical of paternalistic laws (i.e. Prohibition and the War on Drugs seem like a failures that made things worse)
I think it’ll be really hard to regulate this in any kind of meaningful way. This is much harder to regulate than drugs or alcohol.
My guess is that porn and romance novels/movies have a mixture of effects that include “making sexually frustrated young men less likely to rape or be pushy about sex”, but also setting unrealistic expectations that subtly warp relationships.
The era of “human fertility” is really just a blip here – within another couple decades (at most 50 years) after most after AI partners become threatening, I’d bet moderately heavily on us having uploads, and full fledged artificial life being pretty common, and pretty quickly dominating the calculus of what’s at stake. (there might or might not still be biological humans, I’d probably bet against biological humans lasting more than 200 years, but they’ll be a small fraction of sentient life that morally matters)
I think complex relationship games are a “rich” source of fun, and while I think some of the major suffering of breakups or loneliness might be worth changing in the posthuman future, I think the incentive gradient of “fix problems with technology” naturally veers towards Spieglman’s monsters that are simple to the point of IMO not being that morally valuable.
With all that in mind, what is the right thing to here? Hell, I do not know. But rounding this off whatever your preferred simple ideological handle is, is going to massively fail at protecting the future. Figuring out the answer here involves actual math and weighing tradeoffs against each other for real.
Why do you think so? Even if open-source LLMs and other necessary AI models will develop so quickly that real-time interaction and image generation will be possible on desktops, and there will be open-source waifu projects, mobile is still controlled by app stores (Apple and Google Play) where age-based restrictions could be easily imposed. Jailbreaking/unlocking the phone (or installing an apk from Github), connecting it to a local server with AI models, and maintaining the server is so burdensome that I expect 90+% of potential users of easy-to-access AI partner apps would fall off.
This is not to mention that open-source projects developed by hardcore enthusiasts will probably cater for their specific edgy preferences and not appeal to a wide audience. E.g., an open-source project of anime waifus may optimise to trigger particular fetishes rather overall believability of the AI partner and the long-term appeal of the “relationship”. Open-source developers won’t be motivated to optimise the latter, unlike AI partner startups, because their lifetime customer value would directly depend on that.