Or no AI enhancements, since you have the raw photo sensor data, even better you can pick your own AI algorithms to apply to the image.

E.g. from the raw data we know the sensor, we can train a specific AI to enhance the image from the known physical properties of the CCD