Speech recognition technology has gained considerable attention in the nineties and its position in the marketplace is getting stronger. The ultimate goal for speech recognition systems is to be able to converse fluently, without any restrictions, with a machine. Though the fulfilment of this ideal goal seems far away, there are a lot of interesting and important applications which do not require a large vocabulary or where it is not a serious constraint to force the users to speak in an isolated word style. Putting such restrictions on the systems makes them realistic to implement today.
Here, we will review the idea behind speech recognition systems, their problems and key methods. The activity performed in Norway will be described and we will give some examples of successful applications which hopefully can give some ideas about where and how this technology can be utilized in Norway.