Friday, 2 February 2018

SEEN AND NOT SEEN / Escaping the face

'He would see faces in movies, on TV, in magazines, and in books, he thought that some of these faces might be right for him. And through the years, by keeping an ideal facial structure fixed in his mind, or somewhere in the back of his mind, that he might, by force of will, cause his face to approach those of his ideal. The change would be very subtle, it might take ten years or so, gradually his face would change its' shape, a more hooked nose, wider, thinner lips, beady eyes, a larger forehead. He imagined that this was an ability he shared with most other people, they had also moulded their faced according to some ideal. Maybe they imagined that their new face would better suit their personality, or maybe they imagined that their personality would be forced to change to fit the new appearance. This is why first impressions are often correct.'
Talking Heads - Seen and Not Seen (1980)

Psychometrics developed in the 1980s, a process that contemplates your personality and attitude traits through a personality test, called OCEAN. It was originally utilised in terms of optimising your performance within a business. Ocean assigned participants to a sliding scale of personality traits, from liberal to conservative, introverted to extroverted, etc. These could then be tessellated with your co-workers results, into maximum performance groupings.

It then started to filter into advertisement, as it gave the ability for targeted advertising to become even more personal, directly addressing personality, in a more complex way than previous alternatives that established demographics in terms of gender, income, age.  

In the 2000s, Michael Kosinski was studying a PHD in the psychometrics department at Cambridge, and uploaded the OCEAN personality test to an early Facebook. He expected to only get a handful of responses, but it spread and soon he had millions of responses. Suddenly he had the largest set of responses to the OCEAN test, and was able to correlate these with peoples interests, which words they chose to use, how many people they were tagged alongside in photos, etc (1). 

Being able to correlate people's psychological traits with their digital footprint, Kosinski was able to judge your personality from the way you write. For example, from what I've written so far, it says I am not engaged with the outside world and am 44% relaxed. You can try it here (2).

These processes were utilised by the leave campaign for Brexit, and surfaced as a programme called VICS. One of the things VICS offered, was an app for leave canvassers, which suggested certain households to target, based on their likelihood to be responsive, a prediction pieced together from various data, both online and from the electoral register (3). 

With an increase of trackable wireless networks and phone GPS, the data produced by your movement can also be added to the mix, and correlated to your dataset (4). Your personality can be compared to which areas of the city you visit. This is along the lines of UBER's predictive fare charging, which will judge how much you would be willing to spend, based on your location and destination (5). 

Kosinski returned to cause further controversy in 2017 with developments in the field of biometrics, combined with his previous research into psychometrics (6). This time using dating websites as a database of faces, and filtering descriptions through the 'open text prediction' in order to produce results in terms of the OCEAN test, he started to correlate peoples IQ, political leaning and sexuality to facial structure. 

Whilst the new iPhones feature 'face ID', recent developments of CCTV has led to facial recognition that predicts your age, ethnicity, gender and even mood (7) and NEC gaze detection understands objects of interest via multiple people looking in it's direction (applicable to both security and advertisement). On top of our writing style, internet choices, location within the city, and however controversial and scientifically unstable, your face is fast becoming another token within the data whirlpool.