I've thought of this "proof", but it seems lame to me (I'm not sure it is even a proof), I probably should have used other properties of complex numbers to write one.
So,
$z^n=1 \implies \sqrt[n]{z}=\sqrt[n]{1} \implies \sqrt[n]{1}_1 \times \sqrt[n]{1}_2 \times ... \times \sqrt[n]{1}_n = (\sqrt[n]{1})^n = 1$
And we know there are $n$ terms in the product because it is given by the fundamental theorem of algebra (which says that a polynomial of power $n$ has $n$ solutions). And the next step is given by the fact that the product of powers of $x$ is $x$ raised to the power of the sum of the powers of the factors.
This is obviously a homework, so, please don't assume much when replying! :)
No proof is possible. The conjecture is false.
Suppose $a_i$, $i = 1, ..., n$ are the solutions to $z^n - 1 = 0$. Then
$$z^n - 1 = (z - a_1)(z - a_2)...(z-a_n)$$
When you multiply out the brackets, you get a polynomial in $z$ with constant term $$(-1)^n a_1 ... a_n$$
So, by equating coefficients, we have
$$a_1 ... a_n = (-1)^{n+1}$$
Possibly, the question is asking for the product of the non-real solutions to this equation.