I'm considering the $\kappa$-Minkowski space – a certain model from non-commutative geometry which reduces to the usual Minkowski space in the limit $\kappa \rightarrow \infty$. The parameter $\kappa$ is a new fundamental constant of order $\text{length}^{-1}$, which is conventionally taken to be $l_P^{-1}$ where $l_P$ is the Planck's length.
It was proven in hep-th/9405107 that the lowest order polynomial in the non-commutative coordinates invariant under the full Poincare quantum group is
$$ c^2 t^2 - \vec{x}^2 + \frac{3 c t}{\kappa}, $$
which becomes the ordinary Lorentzian invariant interval in the $\kappa \rightarrow \infty$ limit.
According to a straightforward calculation, this invariant polynomial leads to the following dispersion relations for light in vacuum:
$$ p^{2}(E)=\frac{E^{2}}{c^{2}}\left(1-\frac{9}{4}\left(\frac{E}{E_{p}}\right)^{2}+\frac{27}{4}\left(\frac{E}{E_{p}}\right)^{3}+\dots\right), $$
where $E_p$ is the Planck energy.
Note the absense of the 1-st order term $E/E_p$ in the expression above. This reminds me of this PSE question of mine, where an excellent answer by kleingordon@ referenced a paper with a specific experimental bound on the dimensionless coefficient before the 1-st order term. Because the 1-st order term obtained from non-commutative geometry vanishes, that bound is satisfied.
Question: does this mean that $\kappa$-Minkowski has not been experimentally falsified by the gamma ray burst measurements? Or are there any other bounds on subsequent terms that also have to be taken into account?