Observations of RPE disruption and autofluorescence (AF) photobleaching at light levels below the ANSI photochemical maximum permissible exposure (MPE) (Morgan et al., 2008) indicates a demand to modify future light safety standards to protect the retina from harm. To establish safe light exposures, we measured the visible light action spectrum for RPE disruption in an in vivo monkey model with fluorescence adaptive optics retinal imaging. Using this high resolution imaging modality can provide insight into the consequences of light on a cellular level and allow for longitudinal monitoring of retinal changes. The threshold retinal radiant exposures (RRE) for RPE disruption were determined for 4 wavelengths (460, 488, 544, and 594 nm). The anaesthetized macaque retina was exposed to a uniform 0.5° × 0.5° field of view (FOV). Imaging within a 2° × 2° FOV was performed before, immediately after and at 2 week intervals for 10 weeks. At each wavelength, multiple RREs were tested with 4 repetitions each to determine the threshold for RPE disruption. For qualitative analysis, RPE disruption is defined as any detectable change from the pre exposure condition in the cell mosaic in the exposed region relative to the corresponding mosaic in the immediately surrounding area. We have tested several metrics to evaluate the RPE images obtained before and after exposure. The measured action spectrum for photochemical RPE disruption has a shallower slope than the current ANSI photochemical MPE for the same conditions and suggests that longer wavelength light is more hazardous than other measurements would suggest.