Abstract Objectives: To evaluate whether automated external defibrillator (AED) training and AED availability affected the response of volunteer rescuers and performance of cardiopulmonary resuscitation (CPR) in presumed out‐of‐hospital cardiac arrest (OOH‐CA) during the multicenter Public Access Defibrillation Trial. Methods: The Public Access Defibrillation Trial recruited 1,260 facilities in 24 North American regional sites to participate in a trial addressing survival from OOH‐CA when AED training and availability were added to a volunteer‐based emergency response team. Volunteers at each facility were trained to perform either CPR alone (CPR) or CPR in conjunction with AED use (CPR+AED) according to randomized assignments. This study reports the frequency of response and initiation of CPR actions (chest compressions and/or ventilations) by volunteers in the CPR and CPR+AED study groups. Results: A total of 314 presumed OOH‐CA episodes occurred in CPR facilities, and 308 occurred in CPR+AED facilities. The volunteers were matched well for age, gender, and other features. Overall, ventilations (23.1% vs. 13.1%), chest compressions (24.4% vs. 12.1%), and both actions (19.8% vs. 10.5%; all p < 0.05) were more commonly performed in OOH‐CA cases in the CPR+AED group. However, when only OOH‐CA cases with volunteers responding were analyzed, the rates of CPR actions were similar. In the subgroup of CPR+AED cases with a responding volunteer, the AED was turned on for only 47% of cases. Volunteers initiated a CPR action more commonly when the AED was turned on (60.7% vs. 39.3%; p = 0.003). Conclusions: In the Public Access Defibrillation Trial, rates of CPR actions for presumed OOH‐CA victims were low but similar for CPR and CPR+AED responding volunteer rescuers. Factors associated with volunteer response, CPR action initiation, and AED activation warrant further investigation.