@socrates Thank you for sharing! I’m not quite sure if I agree with that policy. I feel like a post that is set to be publicly accessible constitutes public communication.

I would agree that publishing more than just anonymized aggregate data or IDs (like Twitters TOS suggest) is out of the question but analyzing public communication can’t require a specific opt-in from every user. That does not seem feasible to me.

@Kudusch @socrates I feel like Kususch's stance on this is probably more nuanced than is coming across in this thread, but FWIW the Association of Internet Researchers argues, as a scholarly community, that context matters when deciding what should count as "public" for the purpose of research. I recommend reading this post by Hugh Rundle for an idea of how long-time Mastodon users have viewed their contributions and privacy in context. https://www.hughrundle.net/home-invasion/

[Edited for clarity]

@josh @Kudusch Okay the context here is that we told you explicitly not to do it and now you're aware of that expectation

Many of our users came here to *avoid* being included in Cambridge Analytica type situations and we're telling you: No

That's the context

If a person talking in a public place like a coffee shop told you "stop writing down everything I'm saying," what would the right thing to do be?

@socrates @josh I definitely get what you are saying. An explicit opt-out like for the users of scholar.social must be considered. That's basic research ethics.

My background is in disinformation studies, where there are usually very little good-faith actors. Users actively using the fediverse to spread (potentially harmful) misinformation are tough to identify before somehow accessing public communication at large.

@socrates @josh

Maybe implementing something like an opt-in for (non-commercial) research on an account-basis that can be access through the API could be a solution on the long-run.