The big issues in social data science in 2021: a view from Melbourne, Australia (Part 1)
By Anthony McCosker (@ACMcCosker), Diane Sivasubramaniam, Liz Seabrook, Kath Albury (@KathAlbury), Sam Wilson & Jane Farmer (@jane_c_farmer) from Swinburne University Social Innovation Research Institute. This is the first part of 2 articles underlining the key contemporary themes which are relevant for social data science researchers. Part 2 focuses on data user and producer rights, long-term data usability, and ethical Artificial Intelligence.
So much is happening across the many fields influenced by data science. As the impacts of data-driven technologies filter through to more and more parts of society, ideas of public interest, trust, participation and inclusion are coming to the fore. We highlight a few of the most pressing issues and questions for the year ahead from our perspective located as we are — in Melbourne, Australia… but taking a world-view.
What is the public interest in public interest technology?
This term and its synonyms — the public good and common good — seem to be bandied about without much attention to a definition. Although the concept of the public interest has the quality of being familiar and commonplace, it is difficult to define or articulate in a precise or comprehensive way. It seems to be as much about process as outcome. The difficulty of apprehending and agreeing on the meaning of the public interest only increases as a function of the social and technological complexity of a society. It is a social priority to learn to think more clearly about the meaning of the public interest regarding technology and data science, and the pathways through which it might be discovered.
Is AI reinforcing the power of the already powerful?
This question is not new, but it is becoming a big issue for how AI and associated technologies will be more widely used. Cloud services and ‘AutoML’ are being rolled out by the big tech corporations (Microsoft, Amazon etc). As the name suggests, AutoML automates many of the difficult, costly, and time-consuming processes in machine learning like classifying, training, tuning models. But at the same time, it makes the process less transparent. This is handy for organisations or businesses that have limited resources and time, but want to use some AI for what they do. But it embeds disparities in power and accountability directly into AI applications.
Who gets to ‘play’ with data (and who doesn’t)?
It goes without saying that data literacy is uneven at best, the domain of highly trained experts. But plenty of people are self-training and making use of the data available to them. When is data use or manipulation policed?; when is it considered illicit? We’re thinking here of anything from schoolchildren ‘gaming’ Zoom during COVID-19 schooling to the ongoing chaos around the WallStreetbets Redditors’ manipulation of the stock market.
What’s the right balance between the maintenance of civil society and freedom of speech and information?
While the ability of citizens to access information has increased exponentially in recent decades, so has our vulnerability to misinformation. We are therefore faced with a pressing question: how to write legislation that can strike the right balance between the maintenance of civil society, in which misinformation and abuse are curtailed; and freedom of speech and information?
Finally, what is the role of technology developers and providers in ensuring this balance?
The recent controversy surrounding the Facebook news ban in Australia illustrates the tension in this space, and highlights gaps between the way technology companies operate at the edges of national law and public interest.
End of Part 1.
This is the blog for Data & Policy, the partner journal for the Data for Policy conference. You can also find us on Twitter. Here’s instructions for submitting an article to the journal. Part 2 of this article is available here.
