In this work, we utilize the high-fidelity generation abilities of diffusion models to solve blind image restoration tasks, using JPEG artifact removal at high compression levels as an example. We propose an elegant modification of the forward stochastic differential equation of diffusion models to adapt them to restoration tasks and name our method DriftRec. Comparing DriftRec against an $L_2$ regression baseline with the same network architecture and a state-of-the-art technique for JPEG reconstruction, we show that our approach can escape both baselines' tendency to generate blurry images, and recovers the distribution of clean images significantly more faithfully while only requiring a dataset of clean/corrupted image pairs and no knowledge about the corruption operation. By utilizing the idea that the distributions of clean and corrupted images are much closer to each other than to a Gaussian prior, our approach requires only low levels of added noise, and thus needs comparatively few sampling steps even without further optimizations.