Genome-wide association studies (GWAS) have discovered thousands of genetic variants linked with autoimmune disease, and yet the underlying molecular pathways have remained elusive. A key challenge is that >90% of identified GWAS hits are in non-coding genomic regions that makes it difficult to predict their contribution to disease. Analysis of non-coding fine-mapped genetic variants from over 30 different autoimmune traits including common conditions such as lupus, Crohn’s disease, and multiple sclerosis revealed potentially pleiotropic and shared molecular pathways underlying diverse autoimmune diseases. To test the functional relevance of these non-coding risk loci and identify their potential target genes, we coupled a pooled CRISPR activation screen with single-cell RNA-seq and surface protein detection in primary human B cells. Of 762 autoimmune risk loci tested loci, we identified 524 cis-regulatory target gene effects for 378 risk loci with many autoimmune risk loci regulating multiple gene targets. We uncover many examples of complex cis-regulatory landscapes including non-coding regulatory elements that control HLA expression. Our perturbation screen identified many disease-relevant genes, including cytokines and transcription factors, as well as enabled direct measurement of cis-regulatory element transcription in our single-cell transcriptomic dataset. Finally, we combined massive parallel reporter assays and in situ prime editing to discover a gain-of-function variant associated with lupus that controls expression of the transcription factor REL, with potentially widespread and pleiotropic effects on loci that mediate risk for other autoimmune diseases. Our study provides a valuable resource to link cis-regulatory targets and consequences of non-coding autoimmune genetic variation in human B cells and represents a major advance in our understanding of the genetic networks that may drive autoimmune disease.