Privacy and destiny, data and individuality

Pavlov's Sermon

‘Those who cannot remember the past are condemned to repeat it.’ – George Santayana

 

The high importance placed upon ‘the private’, our increased links to one another via the internet, and the power of the same tech to capture and catalogue personal data ensures that debates about the future of privacy are frequent. The concern in Australia about the ABS retaining personal identification data in the Census (for a longer period) being only the latest to cause a panic.

The internet first broadened expectations for privacy as the greater world became accessible from the bedroom, but the undertow now threatens to drag the bedroom (and beyond) back into the public.

A major motivation of ardent defenders of privacy, typified by Edward Snowden, is the concern that the past, or present will be used against them at some stage in the future. For this reason they feel that their freedom to speak and think freely, now, is encroached upon.

These considerations are significant, but so much of the concern about privacy is fixated on variations of this same theme. In this essay I consider another dimension: privacy not to protect what we have done but privacy for the protection of what we will do, who we will be (and who we are).

Behaviorism and privacy

Skinner Box

Let me start with a basic analogy. Rat’s, stats and experiments.

In the 70s Bruce Alexander1The analogy is topical. Bruce Alexander recently released a book on this subject; and last year British journalist Johann Hari, controversial for taking certain ‘liberties’ (that is, poetic license (i.e. making stuff up)) in his journalism, often referred to Rat Park in promoting ‘Chasing the Scream’, a book about addiction and colleagues, began with an experiment which had showed that rats, isolated individually, within sterile test conditions, tended to prefer morphine tainted water over pure water because the morphine triggered a pleasure response, thereby, so the argument goes, this experiment supplied empirical evidence that addiction is biological/chemical: the rats simply prefer morphine over water. Bruce’s team, however, created further experiments which pointed to a different conclusion. They discovered that when the rat environment was modified to be more open and social the rats displayed a preference for the natural water. This was said to be true even of rats which had previously formed a reliance on morphine within the earlier experiment. In other words, rats, otherwise entertained, were less likely to succumb to morphine addiction. In short, if this experiment’s findings are sound, they lend credence to arguments that addiction is more social/environmental than usually assumed.

But what does this have to do with privacy? I use this analogy to draw attention to the importance of data, the effect of the environment on decisions (the environment was the key element that triggered the different (non-morphine) response from rats thereby giving the experimenter some measure of control over how the rats will live) and to draw a parallel between this and the growth of our digital footprints: the much vaunted “big data”. This data, as brutish as it may sometimes be, is definitely a source of information about our human behaviour. And as with researchers and the rats, this knowledge ultimately enables a measure of power over behaviour.

In the following I refer mainly to a commercial use of personal data as a means of influence. This is because the commercial applications of information are clear (and present) ; yet, the content of the essay refers to the phenomenon in general: information > intention?

Privacy and marketing

The focus of advertising psychology through the last century could be summarised as variations of a theme, ‘how can we encourage people, children, and dieters, in general, to buy from us?’. Answers to this question have led to packages are redesigned, super-market layouts altered and so on.

What’s new in the last 10 or so years, is the capacity for companies to gather data more efficiently and tailor advertisements to segments of the population with greater sophistication, such as ‘pregnant mothers’ as in a Forbes report notes of Target, but beyond this, to ‘individualized’ marketing, marketing not to general ‘groups’ but to you in particular.

Common, though crude, examples of this phenomenon can be seen in Facebook-triggered ads based on items searched for elsewhere, or when hotel adverts popup after searching for flights.

This use of our data neither knows much about us as individuals, nor is it overly powerful. It is passive, in that the output (ad) we are presented with on our screens may ‘suggest’ options to choose. Rather like giving the rats the option of morphine water while leaving the back door open, we are free to follow the advertising recommendation or make some other choice entirely.

 

The rat who wouldn’t and ‘active’ advertising

If there was one rat that consistently proved impervious to a morphine addiction while all the others had succumbed, this ‘exceptional’ rat might eventually be excluded from analysis, as an outlier. In a similar way, the personal, unique aspects of individuals have been outliers in calculations of marketeers and ad agencies through history; they were simply not practical to focus upon, nor was data about these traits easy to gather. Besides, lowest-common-denominator rules about human behaviour (or particular niche sub-sets of behaviour) are more effective in terms of economic outcomes. But perhaps there is potential that individual traits no longer need to be collated, (and lost in the mix) with others; that now individuals, in all their many layered complexity, could be treated as sufficient point of data on their own.

This is because, 1) there is more data about us (data which has more details about us): through the increase of GPS location data, digital money transactions, ‘likes’ on Facebook, monitoring our typing habits etc. ; and 2) there is greater interdependence between ourselves and this data (we are now data-creatures) – as a result of the gradual extension of even deeper ways for ‘data’ and ‘reality’ to interweave – such as augmented reality, of which ‘Pokemon Go!’ is only an early example.

In the future, data-analytics would put interested parties in a greater position to actively construct our (digital-organic) environment. And to do so in ways that have an influence on the actions of people similar to scientists and the rats they keep.

Research published by Facebook in 2014, ‘Experimental evidence of massive-scale emotional contagion through social networks’ stands as a testament to this potential. In this study a huge sample (n=698,003) was used in an experiment on emotions (emotional ‘contagion’). Individual’s Facebook news feeds were adjusted so they contained either less negative words, or less positive words. Then the negativity/positivity of each individual’s ‘status updates’ were monitored.

The results were that the increase and the decrease of positivity were correlated with emotional content of status updates (i.e. individuals wrote less positive words in their status updates if they had been ‘primed’ by a decrease of happy vibes) – see the boxplot below. In short, Facebook’s experiment seems, not only to have studied, emotional contagion, but to have produced it. It may actually have affected the emotional lives of individuals ‘participating’, albeit in minute ways.

 

From Facebook research: http://www.pnas.org/content/111/24/8788/F1.expansion.html
Consent? Regarding the tricky issue of consent, Facebook states that “all users agree[d] prior to creating an account on Facebook, constituting informed consent for this research”.

This Facebook experiment was large-scale and randomised – it did not fixate on any individuals in particular (i.e. did not select those individuals who might be more susceptible to manipulation). Yet research along these lines, which fixates on particular individuals, it would seem, is technically possible. And this leads to an elementary point: knowledge of our past (in ever-more intimate details) is knowledge about our future. Or to rework Santayana’s famous phrase, ‘that which knows your past is not condemned to accept you will repeat it’.

If emotions are ‘contagious’ to manipulation by data then there is at least the possibility of connecting moods to other behaviours. Do you behave differently when happy or sad? Are you a ‘comfort eater’? Such knowledge could be ‘in the data’ (because we are, ourselves, enmeshed with that data) and could be deployed to suit any number of different goals.

 

Data and the Self

Whatever one may think about the question of free-will, at the subjective core of human-beings are ideas, perceptions (and reactions). Data about these reactions (i.e. what we do: purchase, not-buy, eat, walk, run, swim) is fundamentally data about us; it is data about how we have acted; it is also, in a sense, data about how we may have acted otherwise. Under which circumstances, have you, faced with a similar choice, acted differentlyIs it possible to have enough information 1) to answer this question; 2) to reconstruct a similar circumstance once more? (Are you a comfort eater? Do you splash out on new shoes  when frustrated in some other area of life?)

Yes, this ‘behaviourism’ commits many but not all, the sins the Humanities have railed against since the mid-20th century – the reduction of human behaviour to ‘external effect’, discarding the inner-life and our ineffable personal sentiments. Arthur Koestler

 

But as data grows, this criticism becomes less to the point. It may be true that I do not know what the ‘chocolate cookies’ mean to you personally. Yet I may know you are sad when you eat them; I may even know why you are sad based on other points of data. We do make ourselves known, in extremely significant ways, through our action.

Access to data about individual behaviour is access to the individual. Even remaining on the saner side of Philip K Dick, the use of data to ‘tweak’ an individual’s behaviour could be indistinguishable from what we regard as freedom. The individual would still be choosing what they want. But it may be the environmental circumstances (a la the rats), the ‘inputs’ which they are presented calculatedly, which is the significant factor behind an action they choose. Just as the rats’ responses to the two environments were ‘natural’ behaviours, and the catalyst for behaviour modification was the changed environment, it seems inevitable that, little by little, human behaviour will become subjected to similar experimentation.

Taking this line of thought to its logical (and absurd) extension raises a world in which our choices are already, in a sense, chosen for us. The rat wants the morphine, the rat wants the water. Raising, too questions about the meaning individuality; and about whether individuals deserve privacy – freedom from data – as a ‘right’. Isn’t it true that there is a dignity in the sense of personal ownership, of knowing ourselves, truly, to ourselves alone? Could we lose this sense and the aura we attach to the individual?

I wonder if we will face a coming ‘loss of faith’ as our activities and our ego are more and more dissected via data and computational power. A loss which may persist until we form a different conception that will give (new) sanctity and meaning to human lives. These are more general remarks, drawing on other on other emergent phenomenon: technology and improving neuroscience.

Leave a Reply

Your email address will not be published. Required fields are marked *

*