Charitably, I think you can read into that a not-unreasonable (if unproven) assertion that there are many lower hanging fruits on the tree that would do endlessly more good for the cause to pick than data centers, and AI at least has the arguable potential upside of alleviating some of those specific burdens-- better health care and less environmental pollution through various improved forms of automation. Or at least when addressing the stub claim of the sort in the GP comment, you should assume these fairly straightforward subclaims pre-emptively and respond to this stronger form of the argument. It saves time, at least if you're going to seek out a discussion it does. Plenty of counter claims to them, but it gets the conversational ball rolling in a productive direction and if the response in turn is less constructive then you also know not to bother any more.