> How about we don't ignore things that can help actual people and leave them worse off just because in theory there are also nicer sounding solutions. I appre

> How about we don't ignore things that can help actual people and leave them worse off just because in theory there are also nicer sounding solutions. I appreciate this approach in general, but it's worth remembering that proposed solutions can feel right without actually being right. Indeed, these proposals are often more dangerous than those that are obviously wrong since they effect undeserved complacency. Implicitly you have assumed that AI companions could actually help the elderly stave off their loneliness, but is there any evidence that this assumption is true? Suppose, for example, that AI companions wouldn't actually help at all, but whatever company develops the software markets it as effective anyway. Some local government might read the advertisement and decide to buy a subscription to their software rather than budget for a community center. Ironically, the net effect would be to make the elderly even lonelier.