25 Apr Data-driven or data-drifted?
The progress made in digital technology has led to ever faster computers and greater memory storage possibilities. Large amounts of data can now be collected, monitored and analysed and new opportunities arise for researchers. So in order to pick up as much information as possible from all over the world on my PhD theme, I visited the conference ‘Big Data and Managing in a Digital Economy’. It was attended by more than 400 researchers and practitioners. In addition to researchers in Human Resources Management, Management Consulting, and Organisational Behaviour, a lot of visitors were data scientist, AI researchers, econometrists, psychometrists and practitioners doing HR Analytics. This wonderful new world of datapoints, metrics, analyses, assessments, and data-driven anything has generated an industry of professionals, tools and practices.
Beautiful. Magnificent. Splendid.
In 2013 David Stillwell, together with Michael Kosinski and Thore Graepel, conducted a survey on the predictability of personality by the likings on Facebook, which is now known as YouAreWhatYouLike (YAWYL). More than a quarter of a million people participated, and (here’s the major scientific leap forward) their Big Five traits of personality were deduced from Facebook data! Now it was possible to target specific groups with special messages like upcoming events for the outgoing and advertisements for good locks & hinges or for the strong presidential candidate for the unstable.
At the Big Data conference, Stillwell discussed the ethical issues raised by this kind of projects (and this was only a month after the Cambridge Analytica disclosures). Understanding the psyche of a consumer or voter from their social media behavior can be both useful and harmful.
Nuria Oliver, Director of Research in Data Science at Vodafone, addressed these problems in her key note. The development of this new branch of science & technology is developing to a large extent behind the scenes by only a small group that still understands how it works! The worldwide financial crisis in 2008 had similar characteristics.
Just before departure, I bought a book by Jamie Bartlett about problems that technology is causing for democracy. Bartlett also refers to the YAWYL research, and he warns of the dangers of algorithms in politics. “We used to call this kind of thing propaganda. Now we call it ‘a behavioral approach to persuasive communication with quantifiable results’ ” (p. 83).
Nuria Oliver pleads for using big data for the good (for example for realising the Sustainable Development Goals), and offers six solutions that responsible scientists, analysts and consultants can use. (1) use centric approach, (2) the application of ethical principles, (3) algorithmic transparency, (4) discrimination-aware decision-making, (5) living labs and sandboxes (protected space in which computer programs can operate without disrupting other processes), and (6) multidisciplinary and diverse teams are a must.
Bartlett exceeds Oliver in the number of solutions. He has twenty, and they come down to more or less the same ideas, among others: (1) own your opinion, (2) a new digital ethics, (3) teach critical thinking, (4) new safety nets, (5) safe AI for good, and (6) future accountable governments.
Whether I wanted to leave my e-mail address to be able to send me the receipt, the lady from the bookshop asked. With the phrase “customers who bought this item also bought…” in my head, I said to her: “No, thank you very much.”
Jamie Bartlett ‘The people vs. Tech – How the Internet is killing Democracy (and how we save it). Ebury Press. London. April 2018
Also worth reading on this topic is Cathy O’Neill’s blog ‘MathBabe – Exploring and venting about quantitative issues’