The U.S. Department of Justice Antitrust Division recently announced that it would permit the use of predictive coding to respond to a document demand in connection with the proposed merger of Anheuser-Busch InBev NV and Grupo Modelo SAB. Predictive coding uses computer algorithms to identify responsive documents instead of having each document individually reviewed by attorneys. The process begins with the predictive coding tool creating a randomized subset of the documents to be reviewed, called a “seed set.” Attorneys familiar with the facts then review the seed set for responsiveness, and the results of that review are used to train the computer to identify documents in the larger set that are also potentially responsive. In this case, the use of predictive coding reportedly cost the parties 50 percent less than a traditional document review would have cost.