What is your device structure? Photocurrent is not linear to your light intensity, is it what you meant? This is directly linked to your device structure.
It is not a surprise to find that a transducer (a device converting and energy form into another, including sensors and actuators) has a nonlinear behavior. This can be explained with "the difficulty of realizing a physical device which can function as an efficient transducer". Once efficiency is the main concern, we may stop expecting a linear response as a bonus (which usually cannot be obtained).
Nevertheless, linearity (or wide linear operating region - or wide "linear dynamic range, LDR" specifically for a photodetector) is usually required in many applications (No need to give examples or reasons).
When linearity is an important concern, several actions can be taken:
1) If you are to fabricate the device, different (promising) materials or novel fabrication techniques can be used to obtain a better photodetector structure with wider LDR.
2) If you are to use the photodetector for a measurement or in the design of a system, then you may
i) accept the narrow linear range and try to limit the light power within that range (You can use an amplifier to widen the dynamic range of the electrical quantity)
ii) consider the non-linearities of the device (You should characterize it) and correct the non-linearities in the electronic circuit by applying linearization methods (e.g. by using functions inverse of the nonlinear behavior).
iii) apply "predistortion" to the light by using optical methods (I cannot imagine how this could be performed to light, since optics is out of my profession)
iv) find another device which shows a wider LDR (if available/affordable)
For sure, each of the above actions may be challenging.