[Foundation-l] Personal Image Filter results announced

[Achal Prabhala]

On Monday 05 September 2011 03:53 AM, Kim Bruning wrote:
> On Sun, Sep 04, 2011 at 11:54:44PM +0100, Thomas Dalton wrote:
>>> Yes, exactly! You're smart! :-)
>>>
>>> Now, one definition of censorship is :
>>> * Filtering on the basis of prejudicial labels.
>>>
>>> We're not actually allowed to censor, because censorship is evil.
>>>
>>> If we want to do this, we'll need to figure out a way to make an image
>> filter
>>> which does not use prejudicial labels.
>> Or we just reject that definition as obviously not applicable. If people are
>> choosing for themselves whether to filter and, if so, what on then it
>> clearly isn't censorship.
> [citation needed]
>
> I don't see why it isn't applicable. You have a censorship tool (your
> prejudicial labelling scheme), and you are applying it for its intended
> purpose (albeit mildly).

Hi Kim, I find your discussion of labelling schemes (and the American 
Library Associations guidelines) extremely useful and interesting. Thank 
you for taking the time to explain this carefully. It has helped clear 
up, for me, similar questions to the kind that Sarah and others raised 
on this list earlier.

> I think that's pretty much sufficient to cross the line into actual
> censorship. Even if you can't quite see how right now, AMA probably can
> and has. (I can easily think of some scenarios myself, if you like. In
> fact, I gave some tangential examples on this list today.)
>
> But... even if we can't agree that *that* is actually across the line,
> the same censorship tool can still be used by others for more sinister
> purposes. High quality prejudicial categorization would most certainly
> be a boon for 3rd party censors, in many many ways.
>
> So the options you are advocating are either (arguably) actual
> censorship, or (if we can't agree to that) the enabling of 3rd party
> censorship.
>
> The board themselves in their decision are very careful not to cross
> those lines. My one issue with the board is merely that I think it is
> very hard _not_ to cross the line.
>
> Of course, some people don't see the danger, and blithely cross
> the line anyway. (Thus proving my point for me much better than anything
> I could say myself O:-) )
>
> sincerely,
> 	Kim Bruning
>
> citation:
> 	http://www.ala.org/ala/issuesadvocacy/intfreedom/librarybill/interpretations/labelingrating.cfm


In relation to the ALA link (which is an exemplar of concision and moral 
clarity), I have a few related questions.

1) Would the article rating tool (Good? Useful? Reliable? etc.) or 
indeed any other comparable qualitative rating/ranking (for e.g. GA/ FA 
status) similarly classify as prejudicial labelling? I ask this because 
in the article rating tool, I can see it fitting under the same 
category, but can't see how it would lead to the same results. An 
archive or library would never employ a qualitative rating like we did, 
but it makes sense on a place like Wikipedia, and I guess it's because 
we're not a traditionally constructed archive or library - though very 
similar in some aspects.

2) In relation to labelling and filtering in a system as being discussed 
on this list, how would it be if Wikimedia/Wikipedia did not actually 
facilitate prejudicial labels on images itself but instead built a 
system that allowed individual users to do so in a way only 
viewable/useful to that individual user as well as specific others they 
shared it with? Suppose there was a system like this:

(a) I, as an individual user, can apply whatever labels I like to images 
on Wikipedia, but these labels will be only visible to me when logged in.
(b) Wikimedia/Wikipedia provides a filter I can use, but it will work 
with my values - i.e. I have to be logged in and have trained it to 
respond to my labels, as I applied them.
(c) I can elect to keep my filtering values (i.e. labels) private or public
(d) If public, other users who wish to filter but don't have the 
desire/knowledge to do so can apply my values depending on their 
inclinations and mine (i.e. different kind of filter values, porn, 
violence, religion, etc. and who knows what). That is, a kind of social 
network for filtering Wikipedia.
(e) But in the end, Wikipedia as such - in the general un-logged sense 
and to users who elect not to use labels for filtering in this manner - 
is unaffected by the labels that different users have constructed since 
they have no official place in it; the labels that users apply exist 
solely in that user's space.

My question is, would this be preferable? Or still prejudicial?

Cheers,
Achal

>