Yes.
Even if CSAM would reduce actual abuse (which is a big "IF", last time I looked researched leaned weakly towards that it increased abuse), the argument is essentially a variant of the trolley problem: "we'll let one kid be abused to make the CSAM to save n others". I'm not comfterable with that, especially since alternatives such as cartoons, AI-generated, or just old-fashioned photoshop exist (but whether that's a good thing is debatable, and is rather off-topic here).