Background: Hypoactivation of the fusiform face area (FFA) among persons with an autism spectrum disorder (ASD) is a very well replicated neuroimaging finding; however, a few studies have reported normal levels of activation. Inconsistent findings may be due to ASD sample heterogeneity and/or differences in experimental procedures, with negative studies employing passive face viewing tasks that are not well controlled. Objectives: We studied modulation of the FFA in ASD and typically developing controls (TDC) by directly comparing several different face task paradigms commonly used in the literature. We predicted that FFA hypoactivation would be modulated by attentional and perceptual processing “load”, with larger group differences for more attentionally challenging tasks. Methods: FMRI scanning was done on 19 children with an ASD (ADI & ADOS diagnosed) and 17 age-matched TDCs. Participants completed a series of five face tasks that varied in their processing demands and with stimuli that varied systematically with respect to social-emotional salience. We utilized whole-brain and ROI analyses to investigate main effects of group and task, and their interactions. Results: TDCs showed greater overall FFA activation relative to ASD participants, and more demanding tasks/more salient stimuli showed greater FFA activation than less demanding/salient tasks. Additionally, several parts of the FFA and related social nodes exhibited greater group differences for more engaging tasks. Passive viewing tasks, including tasks which focused attention on the eye region via a crosshair fixation, showed less FFA activation overall and smaller group differences than active tasks. Conclusions: There are likely two factors modulating FFA activity: depth of processing and stimulus salience. Passive tasks appear to not drive the underlying neural systems hard enough to consistently reveal TDC-ASD group differences. Moreover, active tasks are theoretically preferred, since the real life behavioral phenomena of interest involve active social perceptual judgments.