Rheumatoid arthritis (RA) is an autoimmune disease characterized by joint inflammation. In the last few decades, new insights into RA‐specific autoantibodies and B cells have greatly expanded our understanding of the disease. The best‐known autoantibodies in RA—rheumatoid factor (RF) and anti‐citrullinated protein antibodies (ACPA)—are present long before disease onset, and both responses show signs of maturation around the time of the first manifestation of arthritis. A very intriguing characteristic of ACPA is their remarkably high abundance of variable domain glycans. Since these glycans may convey an important selection advantage of citrulline‐reactive B cells, they may be the key to understanding the evolution of the autoimmune response. Recently discovered autoantibodies targeting other posttranslational modifications, such as anti‐carbamylated and anti‐acetylated protein antibodies, appear to be closely related to ACPA, which makes it possible to unite them under the term of anti‐modified protein antibodies (AMPA). Despite the many insights gained about these autoantibodies, it is unclear whether they are pathogenic or play a causal role in disease development. Autoreactive B cells from which the autoantibodies originate have also received attention as perhaps more likely disease culprits. The development of autoreactive B cells in RA largely depends on the interaction with T cells in which HLA “shared epitope” and HLA DERAA may play an important role. Recent technological advances made it possible to identify and characterize citrulline‐reactive B cells and acquire ACPA monoclonal antibodies, which are providing valuable insights and help to understand the nature of the autoimmune response underlying RA. In this review, we summarize what is currently known about the role of autoantibodies and autoreactive B cells in RA and we discuss the most prominent hypotheses aiming to explain the origins and the evolution of autoimmunity in RA.