Insufficient data for a meaningful answer. We don't really have a clear grasp of the relative benefits and harm yet.
Automobiles offer enormous gains in convenience and productivity, but at the cost of a non-negligible mortality rate as well as environmental impact. Society deems this a net-positive.
The two major unknowns with "AI companions" are: can people be trusted with this level of autonomy, and is the massive centralization of personal data going to result in abuse and exploitation.
I doubt any amount of discourse can answer questions of "how much", but perhaps the conversation can anticipate potential harm and "invent the traffic light" before the car crash equivalent becomes common place.
To do that, we have to first answer: what are the characteristics of the failure-mode of an AI companion?
Automobiles offer enormous gains in convenience and productivity, but at the cost of a non-negligible mortality rate as well as environmental impact. Society deems this a net-positive.
The two major unknowns with "AI companions" are: can people be trusted with this level of autonomy, and is the massive centralization of personal data going to result in abuse and exploitation.
I doubt any amount of discourse can answer questions of "how much", but perhaps the conversation can anticipate potential harm and "invent the traffic light" before the car crash equivalent becomes common place.
To do that, we have to first answer: what are the characteristics of the failure-mode of an AI companion?