Mathematica is not amused :
$$
\frac{(\left|t
\right|+t)\text{Ai}\left(-
\frac{\left|t
\right|}{2\sqrt[3]{3}\pi
}
\right)+(\left|t
\right|-t)\text{Ai}\left(
\frac{\left|t
\right|}{2\sqrt[3]{3}
\pi}
\right)}{2\sqrt[3]{3}\sqrt{2\pi}\left|t
\right|}
$$
– ZubzubJun 06 '17 at 14:42
Off hand I wouldn't expect the Fourier transform to be expressible as a function, since your function is not $L^1$. (It blows up for large negative $x$) The result will likely be a distribution as with the Fourier transform for $e^x$, but it would be hard to show that using the integral definition of the transform.
– JoelJun 06 '17 at 14:44