We describe a method for the automatic quality assessment of multi-slice 2D MRIs with FLAIR, T2, or PD contrast. Quality measures are derived by analyzing the air background, which can be used to sensitively detect typical motion artifacts. Validated on 422 brain datasets, quality indices agree with expert ratings with high sensitivity and specificity (>85%). This technique can greatly improve clinical workflow, suggesting when to repeat low-quality scans, thus increasing chances for higher diagnosis accuracy.