It’s a negative externality.
We can argue about the magnitude and the details, but the basic fact remains.
There are ways of dealing with negative externalities. Some work better than others. The details do matter a lot. And we definitely need better ways of tackling problems like this, especially when the cost is less immediate. The more diffuse, temporally removed, downstream, hidden, or controversial it is, the harder it is to get people to take the problem seriously. Let alone actually do anything about it.
We can take on these challenges. Or we can largely ignore them / throw our hands in the air, and watch the consequences unfold.
As much as we are able to, I’d like to try the former. Computing can still move forward and innovate at a rapid rate.