Yeah, it is kinda both in general. Though in this case, the math about this is well-defined: it's possible to increase a percentage either with addition or multiplication and both of those can make sense, just the words we would use to describe them are the same so it ends up ambiguous when you try going from math to English or vice versa.

But the fact that switching between communication language and a formal language/system like math isn't clear cut does throw a bit of a wrench in the "math doesn't lie". It's pretty well-established that statistics can be made to imply many different things, even contradictory things, depending on how they are measured and communicated.

This can apply to science more generally, too, because the scientific process depends on hypotheses expressed in communication language, experiments that rely on interpretation of the hypothesis, and conclusions that add another layer of interpretation on the whole thing. Science doesn't lie but humans can make mistakes when trying to do science. And it's also pretty well established that science media can often claim things that even the scientists it's trying to report on will disagree strongly with.

Though I will clarify that the "both" part is just on the translation. Formal systems like math are intended to be explicit about what they say. If you prove something in math, it's as true as anything else is in that system, assuming you didn't make a mistake in the proof.

Though even in a formal system, not everything that is true is provable, and it is still possible to express paradoxes (though I'd be surprised if it was possible to prove a paradox... And it would break the system if you could).