Noting that most young troops had spent a lot of their youth playing computer games, the U.S. Army used a current adventure game (Unreal Tournament) and existing speech recognition software, to create a game that taught troops how to speak enough Arabic to be a big help in a combat zone. The Tactical Language Training Project (TLTP) used the game engine for the Unreal Tournament game to create a system that presented users (the troops) with realistic situations, where they were talking to Arabs. If they used the wrong word or phrase, or mispronounced too badly, the photo-realistic character on the screen (what gamers call an NPC, or Non-Player Character) would give realistic body, and verbal, language responses. Game engines are the tool kit versions of commercial games, which game developers often put together to sell to other game developers. This is why many different games look, to the experienced eye, somewhat familiar. With a game engine, you can make a new game that looks completely different from the original that the game engine was derived from. But in most cases, developers will buy a game engine that best suits their needs, and save millions of dollars in development costs (even though buying a game engine can cost over half a million dollars.)

Initial versions of TLTP, which teaches the Lebanese dialect of Arabic, were well received. An Iraqi Arabic version is now being development. The many dialects of Arabic are often incomprehensible to speakers of other dialects. And each dialect is accompanied by a different set of body language and vocabulary. English speakers can experience similar differences if they go to rural parts of, say, Britain or Australia and strike up a conversation. Takes you a while to get over the shock and start communicating. Same thing happens to Brits or Australians who go to Maine, West Virginia or Mississippi and speak with locals.