As has been observed, one approach to this problem is to solve the equations over the complex numbers, and then check directly that $a^3+b^3=0$ in all cases. However, it's both useful and more general to observe that the conclusion can be reached without actually solving the equations: since $a^2+b^2=a b$, we have
$(*)\ \ a^3+b^3=(a+b)(a^2-a b + b^2)=0.$
The problem makes sense over any field -- in fact in any commutative ring with $1$ (this condition is only needed so that "$4$" makes sense.) In addition to being simple, an advantage of the solution $(*)$ is that it works regardless of the setting. This is relevant, since the problem did not specify a domain for $a$ and $b$, so a solution based on calculating the roots in a particular context involves an inherent assumption.
For example, in the field ${\Bbb F}_{37}$ (i.e., the integers mod $37$), the equations have the four solutions $(a,b)$=$(9,21)$, $(16,28)$, $(21,9)$, and $(28,16)$. You can check that $a^3+b^3=0$ in each case, but you're back to square one when faced with ${\Bbb F}_{61}$.