Over the past few weeks my inbox has seen a steady stream of emails on the topic of big data and AI research ethics. From the NSF-funded PERVADE study to industry initiatives to roundtables and discussions at the Association of Internet Researchers annual meeting in Tartu, Estonia this week, it would seem data ethics has hit its stride. Yet, should we have hope that there will actually be change, or is it just too late for big data ethics?

As I’ve written again and again and again and again and again on these pages, the tide in the world of big data and AI research seems to have turned decidedly against the notion of outside ethical review. Instead of asking what questions we should be asking that would better human society, data scientists today all-too-often ask what questions are possible with the data and tools at hand and especially what questions would generate the most attention (and hence publication prestige and grant funding).

Even while academia publicly promotes a narrative of soul searching and a return to focusing on what data scientists should do, rather than what they can do, the reality is little changed. Studies generate public outrage and statements of editorial concern, only for the journal and university to turn around after things have quieted down and say they would do little differently. Professional societies that have become synonymous with extensive full IRB review and informed consent welcome into their own journal papers that have neither. Prominent researchers tout…

Read the rest of this discussion piece