Well it’s nearly time for the GSM conference in Barcelona now renamed as the Mobile World Congress. For those of you thinking to attend it’s from the 14th to the 17th of February.
I would have to say it has been a particularly interesting month with conversations wandering into the realm of science fiction. It all started with the mobile phone now an essential part of everyday life but which might have been looked on by many as ‘sci-fi’ back in the 70’s or even 80’s. In fact I even know a few people who didn’t expect it to take off even in the 90’s. So we started off imagining what phones would look like in 50 years time, same sort of thing really but with a more modern fashion statement, perhaps some snazzy wrist band and of course speech recognition and all that, there was no need to press buttons or even play around with touch screens.
Now here comes the first run up against biometrics, do we believe that in 50 years time that our electronic gizmos are going to have near perfect speech recognition? I think we do and in my snapshot of family, friends and colleagues this was not seriously in doubt. I would just mention that people have been actively working on this for the last 30 years and in various ways for at least the 20 years before that. So in the last 50 years we haven’t got there, so what’s going to make it happen in the next 50 years?It is the advances in technology, we are moving much faster than we have ever moved before and at the end of the day there is no fundamental breach in the law of physics. Starting at home we often have this conversation, if it can happen it will happen and if people realise they need it then it just comes a bit faster. So from the novice side of the counter, will speech recognition be perfect (i.e. without errors) in 50 years time. Well again I think we all agree that it won’t be perfect but near it and maybe just 1% error or less. But as my friendly bank manager used to remind me, if you take instructions from 100,000 people in a day that means on average 1000 people are going to have a problem! This was when we wanted to use finger prints for authentication at an ATM.
Now the thing is that this may not matter, in practice the English language has enormous redundancy. There are many examples but here is one,
Aoccdrnig to rscheearch at an Elingsh uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod are, olny taht the frist and lsat ltteres are at the rghit pcleas. The rset can be a toatl mses and you can sitll raed it wouthit a porbelm. Tihs is bcuseae we do not raed ervey lteter by ilstef, but the wrod as a wlohe.
And so to our speech recognition, if we can start handling this form of redundancy in spoken context then why not 100% for comprehension and that’s all that really matters.
So then we move to identity, a fundamental necessity for payments.
In 50 years time we are not going to have smart cards and the like. It’s all going to be in the phone and then we just need identity on the assumption that our money is in some form of a bank account. Now do we believe that will be true in 50 years time? Anyway assuming in the sci-fi world we need to prove our identity in order to get our ration of kwala powder, how do we do it?
Back to biometrics, not speech recognition this time but voice or speaker recognition. This is a totally different problem to the speech recognition that we referred to earlier. I remember once at a seminar hearing the words of wisdom from one of the leading luminaries. I won’t name him because he might be embarrassed but anyway he said that biometrics can only ever be a compromise because the human body is dynamic, it is constantly changing and therefore our biometrics are also changing. Unless you can update a person’s biometric every hour or so then you are likely to have additional errors to the intrinsic measurement error that you will get whether you like it or not.
In the world of sci-fi you can put your hand or finger on the plate and bingo in you go. I guess it’s going to happen in 50 years time, I’m just not sure how?
See you in Barcelona,
Patsy.
