Past, present and future of speech technologies in translation: life beyond the keyboard

The idea of speaking to and with computers is as old as the idea of computers themselves. Today, after more than eight long decades of research and development in computing and natural language processing, such an idea is far from science fiction: systems that allow humans and computers to interact directly through speech are increasingly becoming part of our daily lives and are transforming the nature of human-computer interaction. Speech technologies have reached a point of maturity to be useful and effective in several domains. They can notably be found in customer and tech support services; virtual assistants; in-vehicle navigation systems; as well as the operating systems of personal computers, smartphones, and tablets. In translation, certain researchers, trainers, and practitioners are increasingly showing great interest in the use of speech technologies for translation purposes. Recent work has shown that both speech recognition and synthesis positively influence the output quality, language professionals’ productivity, and workspace ergonomics associated with translation, revision, machine translation post-editing, audiovisual translation, as well as interpreting processes. This paper presents a brief history of speech technologies in translation and reports on a panel held within the framework of the 44th Translating and the Computer (TC44) conference. The panel members represented both academic and international organizations and discussed some of the myths, successes, challenges, and opportunities of working with speech recognition and synthesis in translation research, teaching, and practice.