Why do we need data centers?

The debate over data centers is persistently oversimplified. Some say we need data centers to beat our international competitors or to generate jobs; others say we don’t need them enough to justify the huge environmental impacts, or maybe we do, but not here. I say we need to look more closely at what we ostensibly need them for.

Data centers serve primarily two growing (and closely related) needs: more artificial intelligence computing power, for purposes ranging from medical research to more convincingly “human” search and chat responses, and massive data collection and matching on the public’s buying and other habits, to enable ever more effective, targeted marketing.

But no one seems to know, when they protest the construction of a new facility, what it is actually supposed to do. (It is ironic that one thing AI and big data can do well is better analyze the data on climate change — while they contribute heavily to it.)

I suggest that we have little need for data centers to run AI chats that replace human customer support and sales people. I also suggest that we actively need to not grow data centers to collect detailed information on people so as to better track them. Nor do we need data centers to flood and thus sway product reviews or political polls or to generate the slop that passes for news and entertainment.

Finally, I suggest that we would be better off without data centers that enable people to retreat into their computers and pretend they have a meaningful social life.

There may be good reasons to say “not here, but yes, this data center should be built” — but in order to make reasonable decisions, we need transparency that we aren’t getting. Not all growth is beneficial — or even acceptable.

Steve Gruenwald, Schaumburg