This is a GCC bug. The lookup for the name on the left of :: "considers only namespaces, types, and templates whose specializations are types" according to the C++ standard, so following the normal unqualified lookup rules it should find both the namespace ::N1 and the struct N2::N1.

Modified example showing why lookup to the left of :: should consider types:

namespace N1 {
const char* f();
}

namespace N2 {
struct X {
static const char* f();
};
using N1 = X;
}

using namespace N2;

using N1::f; // Clang: ambiguous, GCC: N1::f (OK)
struct Q : N2::N1 {
using N1::f; // Clang: ambiguous, GCC: N1::f (error)
};

"When looking up a namespace-name in a using-directive, only namespace names are considered, see [basic.lookup.udir]. — end note" - http://eel.is/c++draft/namespace.udir

So shouldn't the lookup for the name on the left of the :: scope resolution operator when used in the context of a using-declaration 'using ns_name::name;' look up only namespace names?

Ref:
using ns_name::name; (6) - https://en.cppreference.com/w/cpp/language/namespace

6. using-declaration: makes the symbol name from the namespace ns_name accessible for unqualified lookup as if declared in the same class scope, block scope, or namespace as where this using-declaration appears.

"When looking up a namespace-name in a using-directive, only namespace names
are considered, see [basic.lookup.udir]. — end note"

A using-directive is a declaration of the form

using namespace NS;

The case in this bug report is a using-declaration, not a using-directive, so that rule doesn't apply.

GCC bug filed as https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94227

The case in this bug report is a using-declaration, not a using-directive. Agreed.

One last question:
Shouldn't the left-most identifier in the nested-name-specifier in the using declaration always be a namespace name [or empty to refer to the global namespace]? Since a class name cannot be used here, shouldn't it be unambiguous to the compiler?

////////////////////////////////////////////////////////////////////////////
struct N1 {
static const char* f() {
return "StructN1::f()";
}
};

using N1::f;

#include<stdio.h>
int main()
{
printf("%s\n", f());
return 0;
}
////////////////////////////////////////////////////////////////////////////
Compilation Error on CLANG:

...
prog.cc:7:11: error: using declaration cannot refer to class member
using N1::f;
~~~~^
...
////////////////////////////////////////////////////////////////////////////

Refs:
http://eel.is/c++draft/namespace.memdef

"using typename(optional) nested-name-specifier unqualified-id ; (until C++17)"
nested-name-specifier - a sequence of names and scope resolution operators ::, ending with a scope resolution operator. A single :: refers to the global namespace.
https://en.cppreference.com/w/cpp/language/namespace#Using-declarations

Shouldn't the left-most identifier in the nested-name-specifier in the using
declaration always be a namespace name [or empty to refer to the global namespace]?
Since a class name cannot be used here, shouldn't it be unambiguous to the compiler?

A class name can be used as the left-most identifier in a using-declaration. For example:

struct A { int x; };
struct B : A { using A::x; };

struct C { enum { e }; };
using C::e; // ok since C++20