The A-Z of AI

Speech recognition

Technology that helps AI to understand human speech.

Speech recognition systems are trained to recognize what human beings are saying.

A tourist in an auto rickshaw is using his smartphone to translate a sentence for his driver. Speech bubbles above them translate the words "I know a shortcut."

These systems form the backbone of everything from dictation software to language translation tools to voice-activated smart speakers.

While machines may recognize speech, this does not mean they understand it the way humans do.

Human beings can understand language, even out of context or if the sentence comes out jumbled. Machines find this harder.

"Natural language processing" is a recent advancement in speech recognition that enables AI to draw on grammatical rules and analyze live speech in order to better understand the complex ways people speak. This helps AI systems to grasp how something such as tone or humor might alter the meaning of a sentence.

These technologies are steadily evolving to understand not just what we say, but what we mean. With AI design teams continuously looking for ways to build more nuance into their systems, people are having more seamless, natural interactions with AI than ever.