A Blog by Jonathan Low


Jul 22, 2016

How Will We Manage Our Digital Selves? The Battle Over Personality and Prediction

Through our use of the internet we are in the process of creating a digital self. Discerning consumers have become aware of the fact that every click download and purchase is data to built into a predictive model that can help marketers and government officials build a profile of us, and an increasingly accurate one.

What is becoming clear is that we are also developing an online personality whose shape can be both identified and accessed. The power of this information in helping to shape decision-making - our own and those of enterprises hoping to sell us anything from a laundry detergent to a political candidate.

One of the most important coming economic, ethical and personal battles will be over the degree to which individuals and those enterprises agree to share management of that data. Europe is already insisting upon a right to assess the personal impact of algorithmic use as well as a right to forget. But it is what data are remembered - by whom and for what purpose - that will most profoundly influence public and private futures. JL

Dave Winsborough and colleagues report in The Guardian:

Abilities to extract personality from our digital behavior, predicting how we’ll behave, is close. Researchers (have) used Facebook likes to identify personality, gender, political affiliation, sexual orientation and religious affiliation. While the algorithms (and) profit models are unfolding, our digital personalities are accessible to anyone - more revealing than a conversation, and more accurate than our own hopes and desires. Others (may) know us better than we know ourselves, and use that knowledge to take our agency. (Or) our digital selves allow the power of insight to be harnessed for our own good.
It used to be that only very important things were written down. The deeds of kings, the titles to land and the stories of the saints. For the rest of us, the most that we could expect was for our lives to be noted on tax surveys and tombstones.
In contrast, almost everything we do today leaves a digital trace. Cashless payments, GPS tracking, social media and other applications allow a seamless digital trace of what we do, with whom and where. As the ubiquity of these tools increases exponentially, soon our whole lives will in principle (if not in fact) be archived in minute detail, with increasingly limited ability to opt out. What is a boon for some future historians should cause concerns for us.
What are we to make of this digital footprint? How should we feel about such levels of transparency and archiving of our lives? Luckily for us, the fragmentation of our tools, and the changing fashions between applications and formats means that there isn’t a master algorithm that relates all of these information into a seamless whole. Short of becoming an object of concern of intelligence agencies or hackers, there is, to our knowledge, currently no one with the skills (or motivation) to try and unite our fragmentary digital traces into a coherent whole, a digital portrait, or a digital twin, containing our thoughts, emotions and actions.
However, all of this may change if the marriage of the science of personality and our digital behaviour becomes a fact of life. Forget the long personality quizzes and cookie-cutter personality types. Abilities to extract personality from our digital behaviour and infer who we are, predicting how we’ll behave, is much closer than we think. The implications of this advance could be momentous, for personality is the glue that could unite our fragmented digital traces into rich portraits of ourselves, allowing those who hold a digital fragment to infer us whole and entire. For example, researchers at the University of Cambridge successfully used Facebook likes to identify not only personality but gender, political affiliation, sexual orientation and religious affiliation. The same has been done the same on Twitter, and with an individual’s music preferences.
Why is ability to extract personality such a big deal? Consider an internet bookstore that had your basic personal information and the history of your book purchases and reviews. Up to now, this store could use this data to infer which other books you’d like. Now imagine a personality algorithm applied to this data. Suddenly the internet bookstore is no longer a bookstore. It could infer your dating and vacation preferences, likelihood of having a car accident, or your voting preferences. It may do this imperfectly, of course, but is that any different from how you know yourself?

Importance of personality

Who is Donald Trump really? Is he putting on an act, or does he really think that Mexico is sending the US criminals and rapists? Is he really angry all the time, or does he change when off stage and chill out? Does he truly look in the mirror and think his hair is looking good? For millions of people in the US, (and billions more around the globe who can only watch from the sidelines) answers to these questions are consequential – they will underpin judgments about whether to place trust in Trump and vote for him, or not.
Although the science of trying to peer into other people’s psyches can be traced back to Sir Francis Galton in the late 1880s, it was the Germans after the first world war who turbocharged it. Reasoning that poor leadership was a key factor in their losing the war, they created formal, structured tests to assess leadership potential. So effective was this method that the US Office of Strategic Services (the forerunner of the CIA) lifted their approach to select espionage recruits.Explaining the degree to which people differ from each other produced the big five set of personality attributes in the late 1980s, a standard approach which still dominates research and practical applications. Numerous studies have shown that this model has tremendous power to predict many real world outcomes (pdf), including health, wealth, extramarital affairs, theft at work, job performance and career success. Use of these tools to assess candidates for jobs has surged over the past few years as companies have sought to better match candidates with roles, to pinpoint characteristics of higher performers, or to provide insights about the likely derailment paths of aspiring leaders.
Anyone interested in seeing how Trump fares under the big five model should read this. And, perhaps, anyone interested in dealing with you or selling you something could soon have an equal ability to read about your own personality and make their own judgements about how to influence you, what to sell to you and perhaps even how to make friends with you.

Age of digital transparency

While the algorithms, profit models and actors in this space are fluid and unfolding, we can assume with some certainty that in the future our digital selves will have personalities that are accessible to anyone who cares to look. These will be more revealing than a conversation with us, and more accurate than our own hopes and desires. It will be, possibly, an age of true digital transparency. What sort of the world will this be? Here we can offer a Manichean thought experiment in the form of two scenarios.
In the first scenario, our digital selves are outside of our control. Our personal data is owned by the hardware and software we use, and sold to the highest bidder. Anonymity and privacy are in demand but very expensive – there is no feasible opt-out clause. In this world our every engagement with the digital world creates food for marketing and social engineering of Orwellian proportions. Here we are talking about the fundamental loss of agency, a world in which our individuality is used as fodder for consumption and manipulation. When Socrates exhorted us to “know thyself”, it is doubtful he ever entertained the possibility that we may live in a world where others know us better than we can know ourselves, and use that knowledge to take over our agency.
In the second scenario, we retain the right to our personal data and carefully monitor who can use it and for what purpose. In this sense, our digital selves become our agents and avatars, revealing our preferences for specific purposes we’ve assigned to them, but without ability to be used obliquely against us. We’ll have as many digital selves as required, ensuring that their fragmentation works for us instead of against us. And, where useful, we’ll use the personality of our digital selves for both self-insight and growth, helping us lead more authentic lives but preventing others from using this information to manipulate or game us.
While reality will likely be a mix of these two scenarios, through individual and collective action we can certainly push towards a world where our digital selves are a tool that enriches and ennobles, and allows the immense power of insight that comes out of knowing ourselves to be harnessed for our own good.


Post a Comment