|
Peter Cochrane's Hard Drive 1996 Sweet nothings and nice ones After all, digital communication is millions of times faster, and more precise, than any analogue human tongue. And yet, in a recent experiment a text-to-speech system successfully interacted with a speech-to-text converter. Both were PC-based with a humanised voice and adaptive digital speech recognition. The remarkable outcome of this experiment was the apparent ease with which these machines could communicate without error. They seemed at least as good as some human subjects, and even in a noisy room they coped well. Most surprising of all, they could also converse over a standard telephone line. Could it be that the utterances of digital machines are so absolute, prescriptive and precise, that they are superior to our own? For us, every analogue utterance of the same sentence, word or sound is subtly different, and to make things worse, we use this variability to convey emotion and meaning. This is further compounded by the fact that often we do not say what we mean, or mean what we say. We also have a habit of creating a wide range of words and sentences meaning the same, or very similar things. In contrast machines employ, and tolerate, far less redundancy. Our variability is a major limiter to concise and accurate communication between ourselves, and worse, with machines. On the other hand it is also a mark of our extraordinary abilities and richness of culture. We are definitely subtle communicators, and in many respects well ahead of machines - for now. So here comes a new and exciting paradigm with mankind and machine ultimately conversing on an apparently common level. We may only be a short way off voice command and interaction with all electrical appliances in preference to the button, knob, keyboard and mouse. Talking to your car, television, radio, home, and computer is increasingly possible. Voice announcements, messaging, e-mail, text-to-speech and speech-to-text are already with us in trial services and some commercial products. Their great advantage is that we can choose and adapt the communication mode to meet our individual skills and requirements. Driving a car while using today's mobile phone, navigator or computer is difficult, dangerous, and illegal. However, talking to such devices is easy and safe. Several machines joining groups of humans in conversation may also seem a strange concept today, but once we can converse with them, it will just happen. I will bring my semi-sentient machine to meetings with me, and so will others. So when two or more machines are present will they talk in a human tongue or binary? For our sake it has to be a human tongue, at least as long as we need to be in the loop of knowing. Interestingly, there will be a new silico-duplicity as the machines converse behind our backs, invisibly shifting megabytes, negotiating, arranging and dealing to our mutual benefit. But their biggest contribution may be to refine our language, and improve our accuracy and efficiency. Peter Cochrane holds the Collier Chair for the Public Understanding of Science & Technology at the University of Bristol. His home page is: |
Telegraph Group Limited endeavours to ensure that the information is correct but does not accept any liability for error or omission.
Users are permitted to copy some material for their personal use, but may not republish any substantial part of the data either on another website or as part of any commercial service without the prior written permission of Telegraph Group Limited.