In this podcast, Vlad Sejnoha, an expert in speech recognition software, describes why recognizing and transcribing spoken language with computers is not as easy as you might think.

Podcast produced by David Levin. Interviews by David Levin and Rachel VanCott. NOVA is produced by WGBH in Boston. Funding for NOVA is provided by ExxonMobil, David H. Koch, the Howard Hughes Medical Institute, the Corporation for Public Broadcasting, and public television viewers.

Voice recognition is just one part of secret surveillance techniques being used by U.S. Intelligence agencies. To learn more about them, visit pbs.org/nova/spyfactory.