@chrisamaphone @cfiesler it's a hard question!
The Fediverse culture seems better aligned with research where people opt in to participate.
That said, I see a case to be made for opt-out research that benefits the Fediverse at large.
As long as it does not burden communities (e.g., overwhelming a server), puts people at risk, or breaks people's expectations.
Then there are also legal questions, e.g., GDPR.
It'd be great to come up with certain norms and agreed upon research practices.
@andresmh i don’t even think there needs to be consensus - the whole point is people can set their own norms as communities within an instance - but it would be great to get a sense for the spectrum of attitudes and documented tools or protocols that respect them
@chrisamaphone agreed! Are you talking about doing research on just one instance? are you planning to gather data of interactions with other instances? For example, hci.social holds tons of data from other instances by virtue of federation.
@andresmh right; i guess that’s part of the question — is it ok to follow the policies of just the server you’re collecting from, or would you need to consider it for all servers whose data lives there
@chrisamaphone @andresmh isn't this well regulated by GDPR (and similar national privacy laws) already?
Aggregated data, as long as privacy-preserving, is ok. Would be interesting to see instances that decide to systematically share the one they produce. This would need to be scrutinised carefully.
Individual data is subject to dedicated consent per each study, and this is no different from commercial social networks. Notice that quotes, even if claimed to be anonymised, are traceable to their authors by search, so not anonymous.
@mapto @andresmh more concretely: does GDPR even say anything about the data being anonymized? if it’s not anonymized on the public internet, why should i need to de-identify it? journalists quote tweets from people without permission all the time, and iirc this is legal. (but not necessarily ethical, of course!)
@mapto @andresmh as @cfiesler has reminded her followers many times: just because it’s legal (or approved by an IRB) doesn’t mean it’s ethical