AI companions are final stage of digital addiction, and lawmakers are taking aim
technologyreview.com
technologyreview.com
Without seeing the evidence this feels like a cynical attempt to make political hay out of a tragedy.
The person whose AI companion forms their most important social connection is already not doing well psychologically. And from what I’ve seen, they only offer salubrious advice.
I personnaly think it's like the relation between some massacre and videogames, where the attackers were already mentaly ill before committing the act.
But I wonder, we know people (even kids) can make a difference between reality and a videogame. But will we be able to make a difference between a chatbot and a real human ?
Gladiator at Law by Pohl and Kornbluth was also pretty good. Computerized jury trials. The "jury box" turned out to be a computer.
No one in their right minds would want such a product.
Seems like the industry is too high on their own control fantasies to discern what people want.
Mental health needs to be addressed before the propensity for someone to be addicted to AI chatbots can develop at all. By the time someone's AI addiction needs to be addressed, something has already gone wrong.
The same goes for tobacco/heroin/gambling/other addictive substances. It's not very different in practice for AI chatbots. It fills a personal void in someone's life. That void is not usually discussed when building up regulatory frameworks to address the symptoms of despair.
Whoever is using chatbots obsessively, is using on others, not on themselves. Not to fill a void, but to play and harass other humans.
"Digital waifus" are old decade stuff, a very narrow market that has very low value. Those people aren't even addicted, it has become a healthy hobbyst culture.
If you are serious about mental health, the real issue right now is the growing "incel" supergroup of targeted "dangerous" personas. That also has nothing to do with addiction. Those kids need real friends, AI companions _will_ be an extra insult to them, not a pax to fill some hypothetical void. They are not lonely and sad (that's the label people put on them), they are desensitized.
Can you elaborate on this, particularly on how we might solve that issue?
The idea of AI companions plays with the notion of the 2020s "incel". A sad boy who can't get a girlfriend and spend time with games and digital toys. There is no other possible audience for a product like this. My assessment is that this interpretation is wrong, and those kids are not sad and lonely.
That "personality" has been shaped and used online for all kinds of purposes. I think deep down they know that and will see the "digital girl" product as an insult.
The fact that this demography uses LLMs is not a tell that they need someone to chat. Those kids are using AI chatbot content on others, not themselves. What they want is an LLM thug or bully agent to retaliate who mocks them.
What they really need though is a friend, who can understand a little bit of what they went through, and that won't try to use them for some shady purpose or mock them.
They will eventually realize that tech is not that friend, and have been manipulating them all along.
Take some time to think about these issues before dismissing my point. I know you will dismiss my point, but take the time to do so. Thinking about humans as humans is important.
I think you greatly underestimate how widespread and serious loneliness has become, if you think that "sad teenage boy who can't get girlfriend" is the only audience for this. I know many people of various ages who have expressed interest in this, and none of them are what anyone would call "incels"
Friendship and community is not as common as it should be for a huge amount of people, and AI companions will appeal to anyone who feels lonely enough.
As I mentioned, my assessment is that "expressed interest" might not be what it really seems to be.
Of course other groups have "interest", but the real market are incels, isn't it? The other demographics are that expressive of a target?
We'll see. I can't change how things go, and companies are stubborn and will develop it anyway in some form or another.
I don't even know what you're trying to say here.
> As I mentioned, my assessment is that "expressed interest" might not be what it really seems to be.
Yes well I'm not sure how it really seems to be to you. You say these groups aren't lonely, but what they really need are friends...?
It's not hard to grasp. An AI bot is not a friend, it is insulting. If I'm right it will be rejected or misused.
You are wrong, in that people will turn to these products.
I am telling you this after discussing such things at lengths with people, including the pros and cons of it, and why such things appeal to them. My conclusion is certainly not that they feel insulted by it. Feel free to dismiss that anecdote as, I don't know, 'market research' or something.
Just because they are constrained from expressing it, it doesn't mean they do not feel insulted. Inability to represent a feeling is not the same as acceptance. A common mistake.
The "market research" I am referring to is the use of statistics and social profiling, not this conversation or any anecdote. I think those methods are broken.
I think it's cute that you are slowly trying to steer the conversation towards this "feeling of loneliness" by selecting what parts of my answers you decide to quote and address.
However, the parts you chose _not_ to address tell more about who's trying to create a convenient narrative than the comments themselves.
As I said, the process of creating kind human beings that are not fools is what I aim for. Definitely not winning some vain debate.
I'm trying to make you understand that maybe there's other valid reasons people may be drawn to AI companions - and if I'm right, your misunderstanding of why people are or are not drawn to such things will make it really hard for you to make any progress in your goal of "creating kind human beings that are not fools"
Anyway, in my experience, most people are in fact kind by nature and not fools. Maybe it's just a matter of perspective and approaching people with a bit more good faith; not assuming they're unkind, or telling them how they really should feel. Maybe if people were willing to see the other side a little bit more, and have a bit of empathy for others' feelings, they'd not feel the need to retreat from the world.
The discussion on whether humans are good or bad by nature is not new. So is the discussion of nature and nurture. Let's assume we're both familiar with that, otherwise the discussion will be dragged to philosophical mud.
"Making kind human beings" is not a goal. It's not a project I am working on. It's the way I framed the problem. That is in fact the most empathic way I could think of. Definitely better than thinking of products.
Desensitization is a social condition, not a sign of unkindness. It has been demonstrated by the "guard" personas in the Stanford Prison Experiment. Incels are not the only current branch of online people presenting that symptom (although they are the only ones that _appear_ to be sad and lonely).
Empathy is very different from talking soft and making good will speeches. Specially in an environment such as the internet that is deprived of many social cues.
No one "wanted" computer gaming to be driven by how to maximize revenue extraction--mostly from addicted whales.
Yet, here we are.
If there is a way to extract money, someone is willing to stoop that low. This is why we are supposed to have laws.
I would say computer gaming had sucessfully removed the money spending bad part of gaming and turned into a real culture. Streaming brought back that bad part.
But that's another topic. It is hard to compare.
Of course we need laws. What I want is laws that work and address the real issues.
To me it just feels super hollow since the companion has no independent life. Simple human interactions like “how was your day” has no meaning in that context. It just echos back what you’re bringing up with little to no autonomous agency