Re: The Dangers of Web Annotation

Reading through the thread, some notes..

An in Genius historic reference, to set the context:

About a year ago, I tried Genius. My (not anonymous) annotation was
removed by Genius in a few hours without any explanation. When I email
asked them to explain, they asked me to tweet the page url with a
(#genius..bla..) hashtag and ask. I imagine that tweeting provided
them study-able secondary id of a new registrant, in-spite of a tweet
with potentially negative content for Genius. I wonder if this means
that blocking (or allowing) of my annotation could then be based on my
profile on another known public platform?


Stain: "Decentralization basically also means it is much harder to
block harassment - as web authors would have more than just a handful
of sites to check."

Stain, This is not obvious for me. So let me create a couple of scenarios.

1. Genius is decentralized, like Diaspora.
The nodes are sharing annotations for buddies across nodes. The users
are not one large community by default. As users associate with
location or subject proximal nodes, it would be easier to block but
harder (hopefully very hard) for the authors to check what is
annotated by users on other nodes. The annotations that are shown on a
page are coming from proximal node (a subscribed to service node, say
Bangalore-Gees).

2. Genius or Hypothesis or Swtr service can be used to create
annotations (*), while the Genius users get to see, via sharing
arrangements, annotations created using Hypothesis or Swtr - thanks to
standardization. If anonymous annotations are not allowed, then
inspite of standardization, annotations from Hypothesis or Swtr are
not shown unless the users are identifiable by Genius. Provided they
are, its not anymore harder to block. (*): created annotations can be
kept in service managed repository or in 3rd party private
repositories (owned by individuals or groups)

Dan: on without consent.
https://hypothes.is/blog/preventing-abuse/

I am thinking that we may need to accommodate identifying the
author(s) of a page, in the model, before consent can be discussed.
Then to let the authors declare to annotation services that present
annotations to the public to respect their wish. If today, Genius or
H!, will only show the annotations to registered members, this will
remain respectful of the authors wish along with the rules that engage
the user with the service. so, "respectfully" without consent.

We have these use cases with (say Swtr like service):
1. a browser plugin, which pulls all public annotations on a page (if
user is registered as part of group-repository then relevant
annotations from the group),
2. a swtr server-add-on, okay-ed by the author (or the page service
provider) - now its upto the "author" to moderate what is seen on the
page,
3. user override, where a person who is visiting the page can opt to
see the annotations made by their buddies or of those provided by
curation services that they have subscribed to.

Ivan:
" I still want to understand what exactly the differences are between
the harassments via annotation systems and harassments via, say,
Facebook."

If Facebook posts can be seen as appearing alongside web browsing,
then I do not see them as very different. But, its not easy to think
of that, as most of the (harassment) activity on FB happens in
troll-able (among friends?) areas. For this comparison, we should be
able to see (all) FB posts as public. FB has been provisioning social
privacy controls more and more. This is how we may envision the future
of Web Annotations - a nice manifestation of *social* semantic web.
Not as world is flat view of what Genius of today is (English speaking
users of Orkut feeling they were harassed by responses in Portugese -
how did that resolve?). While harassment in general is certainly an
issue, we might have a good start by trying to address it in the
context of harassment by "friends" and associates.


*Web* Annotations are much more, they go from tools to address
harassments and then address phishing! (page transformations :) and
then further to go beyond into more inclusive Web for the billions of
new users. Maybe where people-networks with subscriptions and curation
become essential services for collaboration, moderation and trust.
One of the early demos we built here was about enabling a comment
section on the sites that had disabled comments. We see a great future
when it comes to low literacy and content accessibility - the pages
can be transformed (by community) for their target audiences. A
corporate (say a bank :) page in English can be re-rendered into
simple Kannada via annotations (where some annotations are used to
replace the original text on the page in a different language! We had
shutdown of our hosting service as they received complaints from
certain corporations that we are a phishing service. For now, we have
over come this by asking the crawlers [that checking for potential
phishing patterns] to prove that they are humans before them seeing
these alternate pages. But *Web* Annotations need the relevant context
of its social network (web scale - distributed and decentralized)


-dinesh
( SWeeT Web )
janastu.org



On Fri, Apr 8, 2016 at 8:22 PM, Benjamin Young <byoung@bigbluehat.com> wrote:
> So much to cover, discuss, and plan here...
>
> For now, if any of you can spare the time, read this:
> https://themanual.org/read/issues/5/eric-meyer/article
>
> And I encourage you to actually read it. All the way through. Don’t skim. Don't stop after the sad story at the top. Then think about it. Deeply.
>
> Sincerely,
> Benjamin
> --
> http://bigbluehat.com/
> http://linkedin.com/in/benjaminyoung
>
>
> -----Original Message-----
> From: Sarven Capadisli [mailto:info@csarven.ca]
> Sent: Friday, April 8, 2016 8:59 AM
> To: public-annotation@w3.org
> Subject: Re: The Dangers of Web Annotation
>
> On 2016-03-30 04:50, Doug Schepers wrote:
>> Hi, folks–
>>
>> We've focused mostly on 3 things in this group:
>>
>> 1) the annotation model
>> 2) the annotation REST protocol
>> 3) the anchoring mechanism (e.g. FindText API)
>>
>> The charter describes other deliverables [1]; my annotation
>> architecture diagram goes into details on a few more [2].
>>
>> But we haven't really discussed the social implications of Web
>> Annotations, outside of some informal chats. Specifically, we haven't
>> determined notification and curation models, which are critical if Web
>> Annotations are to be used as a social good, rather than an avenue for
>> harrassment; nor have we discussed the idea of opting-in or opting-out
>> of allowing annotations a particular site.
>>
>> There's been an interesting (if disturbing) thread the past few days
>> about how Genius is being used for what could be considered harassment
>> (and for rude comments, at the very mildest). I suggest that we read
>> and discuss the blog post [3], Medium articles [4], tweets [5 – 10],
>> and Github issues [11] that describe this abuse, and try to think
>> about what our role, as technologists and standards folks, can do to
>> help the situation.
>>
>> Ultimately, if Web Annotation does take off as a feature of the Web,
>> these cases will become all too common. And I don't think that
>> scholarly and academic uses will be immune (though the accountability
>> and reputation risk will reduce abuse). And if such abuse continues,
>> it reduces the value and incentive for Web Annotation to succeed at all.
>>
>> I don't want to derail the current push towards Recommendation, but I
>> do think it behooves us to treat this seriously, maybe on this list,
>> or maybe in other forums, such as I Annotate, and to discuss it with
>> the broader community on social media, where they have started the
>> conversation.
>>
>> Thoughts?
>>
>>
>> [1] https://www.w3.org/annotation/charter/
>> [2] http://www.w3.org/annotation/diagrams/annotation-architecture.svg
>> [3]
>> https://ellacydawson.wordpress.com/2016/03/25/how-news-genius-silences
>> -writers/
>>
>> [4]
>> https://medium.com/@glennf/citation-appropriation-and-fair-use-news-ge
>> nius-picks-up-again-where-failures-left-off-d640719a82ab#.exsmdb2l1
>>
>> [5] https://twitter.com/brosandprose/status/713380185001836544
>> [6] https://twitter.com/brosandprose/status/714668474904264706
>> [7] https://twitter.com/TheFriskyFairy/status/714843748199493633
>> [8] https://twitter.com/FeralHomemaking/status/714970696867319810
>> [9] https://twitter.com/krues8dr/status/714999625090994176
>> [10] https://twitter.com/krues8dr/status/714999849205305345
>> [11] https://github.com/opengovfoundation/madison/issues/920
>>
>> Regards–
>> Doug
>>
>>
>
> Doug et al, I completely agree with your points.
>
> To what extent do you think decentralisation and individual control of annotations goes towards mitigating some of these problems?
>
> I don't intend to claim that technology solves all these problems nor do I completely understand all of the ramifications of possible dangers of annotations on the Web. It seems like some of the reason News Genius has these issues is partly due to the centralisation of the service, which doesn't allow article owners to control what is displayed on their content.
>
> -Sarven
> http://csarven.ca/#i
>

Received on Sunday, 10 April 2016 04:25:04 UTC