Consider for example, the axiom scheme F, that's the axiom scheme of first order logic. What it contains are tautologies ? If we consider a tautology as "universally true formula", this seem the case.
But, a tautology doesnt carry any new knowledge, why should we use tautologies (=axiom sheme) ?
Although the OP's query is not put clearly, there is perhaps a relevant distinction worth making.
Back to basics. What we are fundamentally interested in when setting up a logical system is regimenting ways of deriving some propositions from others.
In mathematics, for example, we might be given a bunch of axioms of arithmetic, or axioms of geometry, or axioms of set theory or whatever. We want some deductive apparatus for deriving other things that have to be true if these axioms are accepted as true (or at least, true-for-the-sake-of-argument). A logical system aims to systematise, regiment, this deductive apparatus.
There are various ways of doing this systematisation. A so-called natural deduction system, highlighting the point that what we really care about is what follows from what, regiments everything as a whole bunch of rules of inference (typically rules governing the separate connectives, the quantifiers, identity).
But another way of doing things is to lay down some basic logical truths that can be invoked as a "free go" in any deduction -- whether the deduction is from, say, our arithmetic axioms to an arithmetic conclusion, or from geometric axioms to a geometric conclusion. We thereby trade in having lots of logical rules of inference for having some basic logical truths we can invoke at any time as "free goes" and then fewer rules of inference. (Trivial example: we can trade in the rule from $A \land B$ you can infer $A$ for all the logical truths of the form $(A \land B) \to A$, assuming we have the modus ponens rule to work with.)
Being logical truths, the "free go" principles will be "tautologies" in a broad sense -- in a sense they are contentless, telling us nothing special about some distinctive subject-matter. So these have a different status from the contentful axioms of this or that special theory, arithmetic, geometry, set theory or whatever.
Now, these "free go" principles are of course usually also described as logical axioms. And a system of logic that invokes not just rules of inference but rules-of-inference-plus-propositions-you-can-invoke-as-free-goes-in-any-proof is called an axiomatic system of logic.
This terminology is fine, of course. For the so-called logical axioms share with the axioms of a mathematical theory like arithmetic or whatever a status as starting points, things that can be invoked without further proof in the system. However, to repeat the point, the terminology also masks an important difference: logical axioms do have a different status to the special axioms of particular theories -- "tautological" vs contentful. And it is that difference which was, perhaps, worrying the OP.