[Air-L] Incidental findings policies for internet research

[Jill Walker Rettberg]

Have any of you had to write an incidental findings policy for a research ethics review? The ERC wants this for the ethnographic part of a project where involving fieldwork and interviews with people doing cell phone photography in order to understand everyday use of machine vision (e.g. algorithmic image filters, image recognition algorithms etc). 

The literature I've found on "incidental findings" is mostly medical - if you're doing brain imaging to understand ADHD and discover a brain tumour, that’s an incidental finding and you need to have a policy for what to do. This kind of incidental finding is extremely unlikely for an ethnography of internet photographers. If this is all the ERC means, I can probably just say there is little risk of incidental findings that affect the informants’ medical or psychological well-being. The exception would be if a research participant appeared to be suicidal, which certainly could happen – Terri Senft describes this exact thing in her book Camgirls: Celebrity and Community in the Age of Social Networks (2008). So what is a good way to craft a policy for a researcher who suspects an informant is suicidal? And are there other health/well-being related incidental findings that internet researchers should consider?

More complicatedly, my university research ethics advisor also suggests I should have a policy for what to do if the researcher is looking through a participant’s social media feed or photos on their camera roll with them and sees child pornography or evidence of a crime – or if they see something problematic while doing fieldwork in an online forum or at a meetup. Should this be reported to the police? Or should one protect the informant? And where does one draw the line? The suggestion from my university is "In the informed consent form all informants will be informed about the researcher’s obligations to report issues of misconduct" - but this seems obviously far too broad, "misconduct" can mean almost anything. Perhaps actual crimes should be reported though. Though heck, reporting seeing a photo of someone smoking a joint or something seems rather different from finding child pornography in a feed. The purpose of the research in this case is NOT to understand any kind of crime or "misconduct", but to understand the ways we understand, use and are affected by algorithmic images and machine vision. The AoIR ethics guidelines (https://aoir.org/reports/ethics2.pdf) don’t mention these situations, though they include so many other useful considerations.

I haven't found ANY literature citing reports on “misconduct” or crimes as related to incidental findings, although it's probably a good idea to have a policy if this kind of situation were to crop up. I know ethnographers have thought about these things for ages and written about them too - but maybe not as an "incidental findings policy"? Does anyone here have knowledge they could share?

Jill

Jill Walker Rettberg 
Professor of Digital Culture
University of Bergen
 
http://jilltxt.net