Reports suggesting a pathogenic role of autoantibodies directed against glutamic acid decarboxylase 65 (GAD65Abs) in cerebellar ataxias (CAs) are reviewed, and debatable issues such as internalization of antibodies by neurons and roles of epitopes are discussed. GAD65 is one of two enzymes that catalyze the conversion of glutamate to the inhibitory neurotransmitter gamma‐aminobutyric acid (GABA). A pathogenic role of GAD65Ab in CAs is suggested by in vivo and in vitro studies. (1) Intracerebellar administration of cerebrospinal fluid (CSF) immunoglobulins (IgGs) obtained from GAD65Ab‐positive CA patients impairs cerebellar modulation of motor control in rats. (2) CSF IgGs act on terminals of GABAergic neurons and decrease the release of GABA in cerebellar slices from rats and mice. (3) Absorption of GAD65Ab by recombinant GAD65 diminishes the above effects, and monoclonal human GAD65Ab (b78) mimic the effects of CSF IgGs in vivo and in vitro. Studies using GAD65‐KO mice confirm that the target molecule is GAD65. (4) Notably, the effects of GAD65Ab depend on the epitope specificity of the monoclonal GAD65Ab. Taken together, these results indicate that epitope‐specific GAD65Ab‐induced impairment of GABA release is involved in the pathogenesis of GAD65Ab‐positive CA and support the early detection of GAD65Ab‐associated CA to initiate immunotherapy before irreversible neuronal death in the cerebellum.