dwsinger
commented
2 years ago I think that in each individual case, how best to work out what is for the good of its users will be case-specific.
Perhaps this sentence is not quite emphasizing the right words? I don't think that the antithetical sentence "The web is designed to harm its users" is what we are contrasting here; it's more antithetical to sentences that say we're optimizing for the good of something else (e.g. implementors, corporations, governments, etc.).
Yes, I agree, there can be tensions here within a value and between values. If I hold unpopular views, "being safe" means I want to be able to speak without fear of retribution. But if those views are violent or repressive, expressing them make the web unsafe for others.
I am struggling to see a middle ground here between terse statements, and working out at length some of these tensions.
Value statements are only as useful as their specificity. These two in particular need further refinement to be useful:
"for the good of its users" opens a range of questions from "who defines what is 'good' for a user and how is this measured" to "what if something that is good for one user is bad for another user."
"safe for its users" is equally challenging: What if the thing that makes the web safe for one user makes it unsafe for another user? That is in a very real sense what's happening around some privacy policies and features at present. The privacy and anonymity afforded by some platforms is used to protect political speech for marginalized group, but that same privacy and anonymity is used by others to intimidate, silence, and harass those same marginalized groups.
Considering the "user" in this context is everyone with an internet connection and an internet-capable device, it is necessary to further define these values to clearly state what is meant by "for the good of" and "safe for."
One path forward is to explore alternative approaches to these types of values. Of relevance is Capability Approach, a framework where values are centered on what capabilities they grant and enable in the user and whether those capabilities allow the subject (in our case the "user") to do and be what they value and have reason to value. This will be different for different users in different contexts, and allows for the level of nuance necessary to serve a global population.
In a capability context, what is good for the user is providing them the capabilities to do and be what they have reason to value, something that aligns well with the core intent of the web. It also provides the necessary language to discuss issues where a capability (say absolute anonymity) may allow one group to do and be what they have reason to value at the cost of another group's ability to do and be what they have reason to value. Ie "what is the overall utility of this capability, what harms can be caused by it, and how do we mitigate those harms so the core intention of the capability remains in place without it working against itself."