It is well known that systematic reviews, though remarkably efficient at distilling high quality knowledge from disparate data, are time consuming and highly resource intensive. If we expect reviews to keep pace with the ever increasing amount of available data and if we hope to increase the number of ongoing reviews to cover more topics, we must continue to look for ways to streamline the process. In the fall of 2005, Professor Stan Matwin of the University of Ottawa's Computer Science Department and Peter O'Blenis of TrialStat began investigations into how advanced text processing algorithms and machine learning techniques might be used to provide some level of automation in the systematic review broad screening process. The goal was to determine if a machine learning system could be taught to conduct broad screening on abstracts accurately and assume the role of a second screener. The benefit of such a system would be to reduce the amount of human time required in the screening process by roughly 50%. Since October of 2005, Dr. Matwin and Mr. O'Blenis have been experimenting with different text processing and learning algorithms in an effort to achieve a system that can screen with the same degree of recall and precision as a human screener without requiring excessive machine training time. To date, the research has shown promising results. This presentation will discuss the results achieved with machine-based screening so far, the implications that these results may have on systematic reviews in the future, and the steps required to move this technology from the test lab to live systematic reviews.