Fagot and Paleressompoulle (2009) published an automated learning device for monkeys (ALDM) to test the cognitive functions of nonhuman primates within their social groups, but the efficiency of the ALDM procedure with large groups remains unknown. In the present study, 10 ALDM systems were provided ad lib to a troop of 26 semi-free-ranging baboons that were initially naive with computerized testing. The test program taught baboons to solve two-alternative forced choice (2AFC) and matching-to-sample (MTS) tasks. A million trials were recorded for the group during a period of 85 days (Experiment 1). Their analysis shows that 75% of the baboons participated at high frequencies and quickly learned the 2AFC and MTS tasks. In Experiment 2, we compared the baboons' behavior when the ADLM systems were either accessible or closed. ALDM reduced frequencies of object-directed behaviors, but had no overt consequence on social conflicts. In Experiment 3, we tested the process of the global or local attributes of visual stimuli in MTS-trained baboons in order to illustrate the efficiency of ALDM for behavioral studies requiring complex experimental designs. Altogether, the results of the present study validate the use of ALDM to efficiently test monkeys in large social groups. ALDM has a strong potential for a variety of scientific disciplines, including for biomedical research. Supplemental materials for this article may be downloaded from http://brm.psychonomic-journals.org/content/supplemental.