I don't know how compatible this would be with your idea; but maybe instead of having human (or worse, AI) written data privacy policy that people (users or community owners) would have to read and parse themselves — having a certain number of pre-defined labels indicative of a certain level of privacy.
I would imagine that, unlike the amount of information an app could get from you that app stores have to account for; Fluxer bots have access to a limited amount of information, which could allow us to use a more linear "privacy rating" of sorts, for instance:
Level
Guarantee
0
No data is collected, stored or otherwise used
1
Data used, but not stored
...
...
5
Sells your soul to the devil
Those label could be self-delivered (to avoid the situation where the user has no information whatsoever on the bot) with an asterisk indicating that it has not officially been audited yet; maybe with an in-between community audit.
This has, however, both downsides of:
Giving more work to the Fluxer dev team and;
Having Fluxer Platform AB indirectly endorse 3rd-party bots.
Now that I've written that down, that pretty much rules out the "officially reviewed" thing. I thing community-reviewed could definitely be a thing though.
In both suggestions, I think we do have to find a way to explicitly differentiate self-affirmed privacy policy and audited privacy policies.
i think reducing it all down to a single scale with no room for nuance defeats the point. i could imagine one value might be "shares data with fourth party services" [you, fluxer, bot, + someone else!]. but who that fourth party is, is pretty important? i quite care about whether a bot is bridging my messages to IRC, Minecraft, Discord, or Facebook Messenger.
the whole point of "privacy nutrition labels" is that they're easy to understand and not buried in legalese. if there's ANY perception that we need an even shorter summary, the feature has completely failed at its only job. you Do Not Summarize A Summary.
I think we do have to find a way to explicitly differentiate self-affirmed privacy policy and audited privacy policies.
i agree? it should be clear to a user that these data usage policies are self-reported, and that bots can lie about it. but i don't think auditing needs to be officially done . because like. even if you audit a privacy Policy or nutrition labels with access to the codebase and reviewed the infra in prod, there's no continued assurance that the bot will continue to operate in the same way? they could lie on the audit . like Volkswagen did for car emissions.
maybe there could still be a notion of "Fluxer endorses this bot". i would use that on a self-hosted instance. i want to host Fluxcord for my own Fluxer instance; it'd be neat if i could mark that as an "Official bot. Operated by the Platform. The Platform claims that this bot's privacy information is accurate."
but i don't think Fluxer Platform AB should be auditing third party bots on the fluxer.com instance. i don't think they should be endorsing any third party bots. it would completely go against the point of "this feature should be easy to use and first and foremost serve bots that want to be respectful of user consent".
if some law is passed making it more clear how platforms like Discord and Fluxer must handle third party integrations (say, if Fluxer were forced to audit bots by the EU), then i guess this could be revisited? but given that it's Not A Legal Requirement and my motivation is Not Primarily For Legal Compliance, i see no good reason to take on this additional workload.
Thread
Comment by vicky
- Giving more work to the Fluxer dev team and;
- Having Fluxer Platform AB indirectly endorse 3rd-party bots.
Now that I've written that down, that pretty much rules out the "officially reviewed" thing. I thing community-reviewed could definitely be a thing though. In both suggestions, I think we do have to find a way to explicitly differentiate self-affirmed privacy policy and audited privacy policies.Comment by sodiboo
fluxer.cominstance. i don't think they should be endorsing any third party bots. it would completely go against the point of "this feature should be easy to use and first and foremost serve bots that want to be respectful of user consent". if some law is passed making it more clear how platforms like Discord and Fluxer must handle third party integrations (say, if Fluxer were forced to audit bots by the EU), then i guess this could be revisited? but given that it's Not A Legal Requirement and my motivation is Not Primarily For Legal Compliance, i see no good reason to take on this additional workload.