I was recently asked by a journalist working on a piece on the future of human-computer interfaces -HCI- to answer four questions:
- What other HCI technology is being developed, and by what companies?
- What form will it take and will the future of the human-computer interface be intrusive, or subtle?
- Can we really trust companies like Facebook and Google not to put the output of such an interface to ‘other’ uses?
- How will the interface between human and computer be protected and privacy be assured?
That got me thinking. I have been involved in HCI for many years now and more recently I participated in the writing of the second Symbiotic Autonomous Systems White Paper where we discussed the future of several technologies, one way or another playing a role in the evolution of human to system interfaces and I am involved in the Digital Reality Initiative of the FDC that in a way is about human-system interface.
Human machine interfaces have started to direct a machine to perform a given task, like using a key to operate an engine, flip a switch to start the air conditioning. As machines became more complex they required more variety of commands but at the same time they could also accept more powerful commands, mostly in textual form. The point I am making is that human machine interaction was designed, out of necessity, with the machine in
mind. It did not mirror the way humans interact with one another. This situation has not really changed in the last fifty years, but it is starting to change now thanks to software and artificial intelligence. Voice recognition has made impressive progress in the last few years so much, in fact, that some companies have turned to voice interaction. I have now to talk to my television and to my car. The problem is that I got so used to type (with my remote or a selection knob in the BMW) that I feel awkward talking to a machine. Also, and this is probably the source of uneasiness, I have to learn to talk to the television (and to the car). Basically these new voice interactions are a transposition in voice of written command. And we are not talking in the same way we write (at least in the languages I am familiar with). Besides, the interaction is more of a mimic of one of those robot talking in science fiction movies in the last century.
What is missing today, in spoken interaction, is the flexibility on the machine side, to understand a convoluted sentence and contextualise it to derive a meaning. This requires a higher level of intelligence as well as learning a bit of who we are and what we could be after.
Mind you, we already have the required technology, and there are several demonstrators of fluent human machine voice interaction, but this technology is not -yet- affordable for mass market use, eg in your television or car.
I have no doubt, however, that in the coming years voice interaction will displace keyboards, relegating them to technology museum.
Notice how youngsters are starting to use voice messages instead of written messages, because they are more convenient. I know, voice messaging has nothing to do, basically, with human-computer interaction, but the spreading of voice messages is fostering the adoption of voice interaction with machines. Voice messages are different from a face to face voice interaction. You are losing the “interaction” part and make do with a unidirectional stream of communications.