One of the increasing morass of “throw it all against the wall and hope it sticks” AI companies appears to think they have invented this new concept called “friend”. Well, actually it’s probably the fact that they think they’ve invented a reasonable substitute for the existing concept of friends that’s the unsettling part.

It has a very premium URL - friend.com - which I imagine might have cost more than the amount of revenue the product makes if we’re lucky, although I have my doubts.

Just in case it dies the death I hope it does: it’s advertising a crappy-looking plastic necklace thing that listens to what you say. Then you have to get your phone out so it can send you a text in response at various points in time.

The more useless half of a smart-speaker, kinda? Not according to the makers. It’s your friend. Literal friend. Which makes for a hilarious FAQ with “frequently” asked questions including “When will I receive my friend?” and “How much does friend cost?”.

One percent of me looks at this stuff and thinks maybe, just maybe, there’s a temporary use for this stuff in acute situations, if run for reasons other than idle profit. It’s possible I am just an old man raging at the wind, misguided in an adherence to the appeal-to-nature fallacy.

After all the loneliness epidemic seems to be real, and occasionally deadly. Experiences like those we saw when Replika lobotomised everyone’s girlfriends show that some users do form deep bonds with AI entities. It just doesn’t feel like this is the general, long-term, solution we need. Apart from anything else, this instantiation of the idea seems cumbersome, and if you break it, the company decides to stop supporting it or starts charging an extortionate subscription for it then your friend just died.

We didn’t always have a loneliness epidemic, and it wasn’t because Victorians wandered around town dressed in their top hats and gimmicky plastic necklaces. At the very least, please, please, do and show the actual, rigorous, universally acclaimed research to show that these things do good and do not do harm in the short and long term before foisting them upon an increasingly desperate population. It surely wouldn’t be hard to get funding for that.