An artificial change?

Tue 10 Jun 08

The first commercially viable bionic arm, controlled by the thoughts of the individual to whom it is fitted, is currently featuring in UK news coverage. For you who were watching US or UK TV in the 1970s, the technology envisaged in the Six Million Dollar Man and the Bionic Woman, has now become something of a reality.

Indeed the journey into artificial intelligence has proved an enduring one.

In recent times, cognitive neuro-psychology has been ‘discovering’ that decision- making is not a purely intellectual endeavour. Now it is widely believed that both rational and emotional factors influence our decisions.

Check out the endeavours of www.semaine-project.eu , a multi-million euro project which is seeking to develop emotion recognising computers.

Essentially, the concept is to enable the machine to recognise signs of user frustration, boredom or tiredness, by reading your facial expression and responding accordingly…

Imagine this, feeling somewhat jaded at the end of a work-heavy week, with the large project laid out in front of you still in need of much attention; and all you can think about is how you can get away from it all during the weekend; when you boot up your machine, as usual, and whilst waiting to click on an icon, a cheery voice chirps up with ‘Hi [Helen], how are you feeling today?’

And this is no pre-recorded gimmick. The voice is waiting for an answer. It’s poised to analyse the intonation in your voice; gauge the relative light or heaviness in your breath; and maybe even identify the slightest hint of sarcasm in your verbal response?

What if there is no voice, just the silent scanning of the camera, searching out recognisable emotionally induced facial expressions, as contained within the tell-tale signs of every fold and crease?

The project’s authors accord that the computer’s recognition of expression is only as good as the human information contained within the form of algorithms which allow the computer to ‘learn’ the combination of nuances which, in turn, denote ‘normal’ expressions of human emotions.

However, humans are notoriously unsuccessful at reading expressions and other forms of body language, for example just consider the amount of research available into the unsuccessful detection of the incidence of lying…And what about the art of false smiling? With the computer be able to detect the deflecting act of putting on a brave face, and beseech the user to come clean with her true feelings?

Will the machine be programmed to empathise; and will it ever challenge a stubborn or bored user?
What about the user employed by a large corporation who ‘logs’ a high level of negative emotion. What might be his fate on audit?

And as to the user who logs a complete lack of emotional response…?

Will displays of excessive, inappropriate emotions constitute a breach of the terms of employment? Maybe there will be a test case or two, where the issues of constructive dismissal or the (employer) dreaded discrimination legislation may be put to the ‘test’ in relation to emotional conditions or behaviour?

Perhaps the gathered data will form the basis for a longitudinal study into the emotional life of the workforce? Let’s hope it takes account of the fact that higher (than expected) levels of paranoia have been identified in the general public; and that when being watched we tend to behave differently than when we think we are unobserved.

How long until we routinely attach electrodes to measure our pulse and sensors to measure bodily secretions?

Ok, that might be going too far but, really, can there ever be a true (artificially constructed) reflection of the nature of human emotion?

And are we headed towards the attempted creation of artificial emotion?

Hmm, smacks somewhat of the Stepford Workplace template to me…



What are your thoughts?

helen@formulaforchange.co.uk