In this talk I argue that as data systems are integrated into decision-making and information systems across sectors as well as public and private spaces, more attention must be directed to better understanding data harms. Doing so provides us with a better appreciation of where we are heading and where we may want to change course.
The 'Investigating Data Harms' project involves recording concrete examples of harms already caused by uses of data systems as well as interviews with practitioners across the fields of law, social work, and education who are investigating and challenging the data harms they see on the ground through their lines of work. As argued by Eubanks (2018) and Barocas and Selbst (2014) new big data systems do not treat everyone equally, the already marginalized in society are more likely to be negatively affected by big data practices. The Data Harm Record shows that harms are already happening at individual and societal levels as people are targeted based on perceived vulnerabilities, have their personal information used in ways that disadvantages them, and can be subject to discriminatory treatment and social sorting (unintentionally and intentionally) in ways that affect access to services and opportunities (Citron and Pasquale 2014). Collectively, we are now all too aware of how online information environments can be manipulated (Woolley and Howard 2017). Real-world harms are also caused by poor data quality, data errors, and algorithm and machine bias (Angwin et al. 2016). Interviews with practitioners demonstrate the need to enhance current understandings of 'harm' and for system wide changes. The political and legal challenges people are facing in trying to redress data harms make clear the limitations of current democratic systems. I conclude by highlighting and discussing the range of ideas being put forward about how our democratic systems should change to ensure greater transparency, accountability, and means for citizen intervention.