One of the issues that arises again and again when discussing ethics with data researchers is this: where along the research timeline should ethical investigation be introduced? The obvious answer is at the beginning. Often, however, a post-research ethical audit is conducted that gives projects the veneer of ethical responsibility with no actual impact on the research itself.
Kalev Leetauru writes in Forbes Magazine this month about the ‘ethical wash’ issue and asks if we have come so far with ethics-blind research and development in areas like AI that we can’t now retrofit a human-centred framework that protects human subjects.
More broadly, he cautions that data science is not a discrete discipline any more – it is impacting on all areas of research and bringing its ethics-blind culture into traditional social science inquiry.
‘Perhaps the biggest issue is that in our big data world, disciplinary boundaries are breaking down and fields with long histories of ethical review are being inundated with work from fields with no history of review and indeed active movements against such requirements. Data scientists are becoming a universal discipline applying their methods and data across nearly every imaginable traditional domain. Just a few decades ago a documentary study of a vulnerable population in another country would traditionally be carried out by a trained ethnographer deeply seated in human subjects research culture and informed consent with privacy and subjects protection at the forefront of their minds. Today that study might just as easily be conducted by a set of computer scientists who harvested millions of photographs and highly intimate personal details from afar..’