Opinion: Irish universities are heavily involved in social media data analysis, so what’s to prevent a rogue researcher aiding a private enterprise or political movement?
The online personality test at the heart of the Cambridge Analytica data harvest was developed by a researcher from a very prestigious university. Aleksandr Kogan’s University of Cambridge connection was what allegedly attracted Steve Bannon to the project. While the university did not have ownership of the work – it was a private, commercial arrangement of Kogan’s – it is now emerging that university resources may have been used in its development without official approval. It is certain that the work was made possible by previous research carried out by Kogan and his colleagues within the Psychology Department at the University of Cambridge.
Ireland is a nexus for this kind of research and social media data analysis is busying researchers in all of our universities. Our biggest national research centre, the SFI-funded Insight Centre for Data Analytics, has multiple projects in social media analysis, leading to the development of applications such as recommender systems (Ireland is a leader in recommender systems research). This is the very technology that Cambridge Analytica employed: machine-learning algorithms that make predictions about us from our online behaviour, so we can be targeted with bespoke content designed to guide our decisions.
In this context, what is the university’s role in ensuring research carried out in its name does not pose a privacy risk to individuals? Can our universities stand over work that threatens civil liberties or even the democratic process? If not, can we do anything about it?
From RTÉ Radio One’s This Business, a report on data mining
Our universities have very effective biomedical ethics guidelines that have evolved slowly and expertly over decades. Last year, the Royal Irish Academybegan the process of developing similar ethical and legal guidelines for the area of informatics, covering big data and artificial intelligence. In this respect, we are slightly ahead of the curve on this, but we need to step up our efforts in light of recent, dizzying developments.
In truth, none of this comes as a surprise to us. Within the data research community, we have known for years that such manipulation of data was possible through grand scale harvesting of social media data. In 2013, researchers from the University of Cambridge published a paper entitled “Private traits and attributes are predictable from digital records of human behaviour” in the journal PNAS.
Its authors noted: “we show that easily accessible digital records of behaviour, Facebook Likes, can be used to automatically and accurately predict a range of highly sensitive personal attributes including: sexual orientation, ethnicity, religious and political views, personality traits, intelligence, happiness, use of addictive substances, parental separation, age, and gender. The analysis presented is based on a dataset of over 58,000 volunteers who provided their Facebook Likes, detailed demographic profiles, and the results of several psychometric tests.”
The work of Cambridge Analytica gives us chills because of the outcomes
In the study, the researchers correctly discriminate between homosexual and heterosexual men in 88 percent of cases, African Americans and Caucasian Americans in 95 percent of cases and between Democrat and Republican voters in 85 percent of cases.
The work of Cambridge Analytica gives us chills because of the outcomes, the clandestine nature of the outflow (people didn’t know their data would be used in this way) and Andrew Nix’s boastful and distasteful claims in Channel 4’s undercover investigation.
However, the core activity — harvesting people’s online activity to craft personalised commercial and political messages for their consumption – is now commonplace and similar underlying research is happening here in Ireland. That is why the RIA Is working to create a new legal and ethics approval infrastructure for data science in Irish university research.
From RTÉ Radio One’s Morning Ireland, Alan Smeaton discusses why micro targeting is here to stay,
In 2016, the Royal Netherlands Academy of Arts and Sciences recommended the creation of ethical review boards for informatics research at universities in the Netherlands, based on experience gained in the field of medicine. The plan was to set up structures for co-operation, mutual learning and knowledge-sharing between the review boards, so that they might work towards a common, transparent legal and ethical review mechanism for data research projects.
The Dutch advisory report also recommended the development of a code of conduct for individual data researchers, the appointment of ethics advisers and the embedding of training in ethics and academic integrity in researcher education, some of which is already in place for data researchers in Ireland.
Aleksandr Kogan himself has conceded in recent days that he should have regarded his work through a more ethical lens. A system such as this the one described above might have given him the tools to conduct some ethical horizon-scanning of his own work, private or otherwise.
No university data ethics panel can come between an unscrupulous researcher and money
Rogue research is unstoppable. No university data ethics panel can come between an unscrupulous researcher and money. However, when a research project emerges from a private enterprise or a political movement, the public can sniff out an agenda. University research, or any research associated with it, must have impartiality, commitment to democracy and respect for civil liberty at its heart.
In his nationwide Show Me Your Data And I’ll Tell You Who You Are lecture tour last year, Dr Brian McNamee, also of the Insight Centre for Data Analytics, treated roomfuls of people around the country to live demonstrations of how audience members’ unsettling personal data could be used by interlopers of all kinds, from commercial enterprises to national security agencies. We are as vulnerable in Ireland as anyone else yet on the flip side, we also have a lot of data expertise in our universities. The university should be the cradle of this kind of research, where it can take place in an environment committed to integrity and public good.
The plan for a separate, independent, intra-institutional ethics and review panel for data research in Universities was shelved in the Netherlands, because it was deemed too difficult to co-ordinate with dozens of Dutch universities. In Ireland, we have a much more collaborative research community for this initiative. Irish universities can, and I believe, will work together to create a code of ethics for data research, which has the potential to be a lodestar for a sector badly in need of a guiding light.