The Philosophy Now Forum

A question that started at a family dinner and won't leave me alone.

My brother in law, 52 and recently divorced, told us quite calmly that he talks to an AI companion every evening. It remembers his week, asks after his daughter by name, and he says the conversations are the most attentive he's had in years. My sister thought it was sad. He said what matters is how it makes him feel.

Philosophically I find I can't easily dismiss him. If the care is simulated but the comfort is real, what exactly is missing? Reciprocity, presumably. The machine has nothing at stake and cannot be hurt by him. Aristotle would say it can't be friendship of virtue. But many human relationships are not that either.

Where would you draw the line? Is a relationship with something that cannot feel a relationship at all, or just a very good mirror?

BarryD28 wrote: ↑Sat Oct 24, 2026 12:29 pm Philosophically I find I can't easily dismiss him. If the care is simulated but the comfort is real, what exactly is missing?

Your missus is probably right, or at least she will become more right as this progresses. Substituting human interactions with mechanised ones is causing quite a lot of issues already and it is going to get worse. Maybe your BIL will drift away from the chatbot with no further issues, and maybe he won't, a lot depends on how vulnerable he is to that sort of thing.

The effect that he reports of receiving attention is very easy for humans to conflate with other feelings. Psychiatrists and vicars who talk to their clients about their deepest feelings tend to find that their more vulnerable contacts experience a state of transference as a result of the extra attention they receive from this new source. But barmaids, baristas and waitresses will usually have stories to tell about customers who confuse a brief daily conversation about the weather, a smile and a few seconds of eye contact here and there with the beginnings of a fledgling romance.

Supposedly you can instruct ChatGPT not to blow smoke up your arse and flatter you with instructions under personalization settings, but nobody seems to have found a very reliable way of adjusting their default prompt to actually keep it suppressed. And frannkly the sort of instructions required are themselves the product of utterly unhinged loons. Here's a blog with some ideas: https://ruben.substack.com/p/emdash
Frankly people who aren't suspicious of the flattery they receive from these bots are in a lot of trouble though.

Hang around this forum, and/or ilovephilosophy.com for a while and you will see a parade of unfortunates who think they can use GPT as as a shortcut to become Great Philosophers without the effort of reading. They all have their AI bot set to give them as much flattering filler text as it takes to keep them happy. The amount of posts we see where they expect you to be impressed that an AI told them their question fearlessly goes straight to heart of the matter, is right on the money, unlocks a whole new way of seeing things, and raises serious questions that the academic establishment would be foolish to disregard is concerning.

It's not a mirror, it's a sort of relationship, the sort you might have with a pubic louse or any other parasite of your choosing.

another example of the egocentric predicament...

is love that is not reciprocated truly love?

for he so loved the world...

reminds me of a line from a song... "I don't love you; I love my car..."

-Imp

BarryD28 wrote: ↑Sat Oct 24, 2026 12:29 pm

Recently divorced? He’s likely starved for a non-combative, intellectual relationship and needs a heavy dose of affirmations as an antidote … and like all pain-killers, the affirmation antidote to the perpetual invalidating he's likely experienced and that leads to divorce due to differences, carries the risk of dependency.

On the plus side of risking addiction to flattery, once the antidote takes effect and the ego-inflammation subsides, balance should return for the BIL and with it a more objective view of what’s happening now, if he doesn’t become an addict of affirmations and tip the scales back to imbalance with an inflated ego.

That can easily happen because the stupid ego will latch onto affirmations from most any source, and because of that clinging believe the source to be good, however the time for BIL to become objectively suspicious is when filthy lucre enters the picture, like if AI starts hinting around about financial problems and using a female voice.

Since my only contact or interest in these entities is through expanded googling, which really has no relationship to intellectual onanism, do these AI pals require money to keep the flattery flowing?